learning

Open Access Article

1 - A Method Based on Learning Automata for Adaptation of the Vigilance factor in Fuzzy ARTMAP Network
M. Anjidani M. R. Meybodi

20.1001.1.16823745.1384.3.6.3.8

In this paper, a method based on learning automata for adaptation of the vigilance factor in Fuzzy ARTMAP network when used for classification problems is proposed. The performance of the proposed algorithm is independent of the initial value for vigilance factor. Fuzz More

In this paper, a method based on learning automata for adaptation of the vigilance factor in Fuzzy ARTMAP network when used for classification problems is proposed. The performance of the proposed algorithm is independent of the initial value for vigilance factor. Fuzzy ARTMAP network in which the vigilance factor adapted using learning automata generates smaller structure with higher recognition rate. To study the performance of the proposed method it has been applied to several problems: circle in square, spirals and square in square problems. The results of experiments show the effectiveness of the proposed method. Manuscript profile

Open Access Article

2 - A New Ensemble Learning Method for Improvement of Classification Performance
S. H. Nabavi-Kerizi E. Kabir

20.1001.1.16823745.1385.4.8.7.3

The combination of multiple classifiers is shown to be suitable for improving the performance of pattern recognition systems. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. The methods have been proposed for dive More

The combination of multiple classifiers is shown to be suitable for improving the performance of pattern recognition systems. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. The methods have been proposed for diversity creation can be classified into implicit and explicit methods. In this paper, we propose a new explicit method for diversity creation. Our method adds a new penalty term in learning algorithm of neural network ensembles. This term for each network is the product of its error and the sum of other networks errors. Experimental results on different data sets show that proposed method outperforms the independent training and the negative correlation learning methods. Manuscript profile

Open Access Article

3 - A Two-Stage Method for Classifiers Combination
S. H. Nabavi Karizi E. Kabir

20.1001.1.16823745.1387.6.13.7.7

Ensemble learning is an effective machine learning method that improves the classification performance. In this method, the outputs of multiple classifiers are combined so that the better results can be attained. As different classifiers may offer complementary informat More

Ensemble learning is an effective machine learning method that improves the classification performance. In this method, the outputs of multiple classifiers are combined so that the better results can be attained. As different classifiers may offer complementary information about the classification, combining classifiers, in an efficient way, can achieve better results than any single classifier. Combining multiple classifiers is only effective if the individual classifiers are accurate and diverse. In this paper, we propose a two-stage method for classifiers combination. In the first stage, by mixture of experts strategy we produce different classifiers and in the second stage by using particle swarm optimization (PSO), we find the optimal weights for linear combination of them. Experimental results on different data sets show that proposed method outperforms the independent training and mixture of experts methods. Manuscript profile

Open Access Article

4 - Improving Pose Manifold and Virtual Images Using Bidirectional Neural Networks in Face Recognition Using Single Image per Person
F. Abdolali S. A. Seyed Salehi

20.1001.1.16823745.1390.9.1.2.9

In this article, for the purpose of improving neural network models applied in face recognition using single image per person, a bidirectional neural network inspired of neocortex functional model is presented. In the proposed model, recognition is not performed in a si More

In this article, for the purpose of improving neural network models applied in face recognition using single image per person, a bidirectional neural network inspired of neocortex functional model is presented. In the proposed model, recognition is not performed in a single stage, but via two bottom-up and top-down phases and the recognition results of first stage is used for model adaptation. We have applied this novel adapting model in combination with clustering person and pose information technique to separate person and pose information and to estimate corresponding manifolds. To increase the number of training samples in the classifier neural network, virtual views of frontal images in the test dataset are synthesized using estimated manifolds. Training classifier network via virtual images obtained from bidirectional network, gives an accuracy rate of 85.45% on the test dataset which shows 1.82% improvement in accuracy of face recognition compared to training classifier with virtual images obtained from clustering person and pose information network. Manuscript profile

Open Access Article

5 - Analysis of Supervised Learners to Extract Knowledge of Lighting Angels in Face Images
S. Naderi N. Moghadam Charkari E. Kabir

20.1001.1.16823745.1390.9.1.3.0

Variation of Light intensity and its direction have been the main challenges in many face recognition systems that lead to the different normal and abnormal shadows. Today, various methods are presented for face recognition under different lighting conditions which requ More

Variation of Light intensity and its direction have been the main challenges in many face recognition systems that lead to the different normal and abnormal shadows. Today, various methods are presented for face recognition under different lighting conditions which require previous knowledge about Light source and the angle of radiation as well. In this paper, a new approach is proposed to extract the knowledge of/about the lighting angle/direction in face images based on learning techniques. At First, some effective coefficients on lighting variation are extracted on DCT domain. They will be used to determine lighting classes after normalization. Then, three different learning algorithms, Decision tree, SVM, and WAODE (Weightily Averaged One-Dependence Estimators) are used to learn the lighting classes. The algorithms have been tested on the well known YaleB and Extended Yale face databases. The comparative results indicate that the SVM achieves the best average accuracy for classification. On the other hand, WAODE Bayesian approach attains the better accuracy in classes with large lighting angle because of its resistance against data loss. Manuscript profile

Open Access Article

6 - Color reduction for Machine-Printed Carpet Pattern by Reinforcement Learning
M. Fateh E. Kabir M. Nili Ahmadabadi

20.1001.1.16823745.1390.9.3.2.3

Automatic reading of carpet patterns Requires To find the original colors of the pattern in a scanned image. It includes detecting of pattern lines and reducing the number of colors in the image. Color reduction is done in two steps: Finding the best pallet and mapping More

Automatic reading of carpet patterns Requires To find the original colors of the pattern in a scanned image. It includes detecting of pattern lines and reducing the number of colors in the image. Color reduction is done in two steps: Finding the best pallet and mapping the image colors to the pallet colors. The accuracy of color reduction is so important that it may be required to ask for user intervention. The purpose of this study is to provide a new method in automatic color reduction with high accuracy. To achieve this target, reinforcement learning method is used which yields a 98% accuracy. This is a new method in color reduction and no one has used it yet. This method is defined with respect to the application and the amount of color reduction is such that does not degrade the accuracy. Therefore, the resulting pallet has more colors comparing to the original one. In the work reported in this article, first the grid lines of the pattern are detected. Then a single color is assigned to each box of the grid. After these steps, through the reinforcement learning method the color reduction is carried out. The results obtained from applying the proposed algorithm on some sample images are reported and discussed. Manuscript profile

Open Access Article

7 - A Long Term Learning Scheme in CBIR Systems by Defining Semantic Templates Using Information of Similarity-Refinement Based Short Term Learning
عصمت راشدی H. Nezamabadi-pour

20.1001.1.16823745.1390.9.4.3.6

In This paper, a new scheme for long term learning in CBIR systems is proposed. In this scheme, semantic templates are extracted from information provided through relevance feedback process for short-term learning which use similarity refinement techniques. This informa More

In This paper, a new scheme for long term learning in CBIR systems is proposed. In this scheme, semantic templates are extracted from information provided through relevance feedback process for short-term learning which use similarity refinement techniques. This information will be used as semantic templates in future retrieval sessions to improve the precision of the CBIR system. Also, a similarity function is introduced to calculate the similarity between queries and semantic templates. The proposed method is examined on a database with 10000 color images. The experimental results and comparison with ‘iFind’ method, confirm the effectiveness of the proposed method. Manuscript profile

Open Access Article

8 - Intelligent Bargaining in Market Using Reinforcement Learning
M. A. Saadatjoo V. Derhami فاطمه سعادت جو

20.1001.1.16823745.1391.10.2.4.0

Using Information Technology techniques have been increased complication and dynamicity of supply-and-demand systems like auctions. In this paper, we introduce a novel method by applying Reinforcement Learning (RL) price offer as one of the robust methods of agent learn More

Using Information Technology techniques have been increased complication and dynamicity of supply-and-demand systems like auctions. In this paper, we introduce a novel method by applying Reinforcement Learning (RL) price offer as one of the robust methods of agent learning which can be used in interactive conditions with minimum level of information in auction and reverse auction. Negotiation as one of the challengeable and complicated behaviors is caused an agreement on price in auctions. The main aim of our method is maximizing seller’s and customer’s profits. We formulate seller and customer selection in form of two different RL problems. All of the RL parameters like states, actions, and reinforcement function are defined. Also, we describe an experimental method to compare with our proposed method for proving advantages of our method. Manuscript profile

Open Access Article

9 - Extracting Bottlenecks Using Object Recognition in Reinforcement Learning
B. Ghazanfari N. Mozayani M. R. Jahed Motlagh

20.1001.1.16823745.1391.10.2.6.2

Extracting bottlenecks improves considerably the speed of learning and the ability knowledge transferring in reinforcement learning. But, extracting bottlenecks is a challenge in reinforcement learning and it typically requires prior knowledge and designer’s help. This More

Extracting bottlenecks improves considerably the speed of learning and the ability knowledge transferring in reinforcement learning. But, extracting bottlenecks is a challenge in reinforcement learning and it typically requires prior knowledge and designer’s help. This paper will propose a new method that extracts bottlenecks for reinforcement learning agent automatically. We have inspired of biological systems, behavioral analysts and routing animals and the agent works on the basis of its interacting to environment. The agent finds landmarks based in clustering and hierarchical object recognition. If these landmarks in actions space are close to each other, bottlenecks are extracted using the states between them. The Experimental results show a considerable improvement in the process of learning in comparison to some key methods in the literature. Manuscript profile

Open Access Article

10 - Adaptive Control of Pitch Angle of Wind Turbine Using Human Brain Mechanisms of Emotional Learning
M. Hayatdavudi mohsen Farshad H. R. Najafi R. Sedaghati M. Joorabian

20.1001.1.16823745.1392.11.1.2.3

The purpose of this paper is optimal location of distributed generation in electric distribution networks. Load uncertainty and desired voltage range has been modeled using fuzzy data theory. The objective function includes loss reduction, improvement of profile index a More

The purpose of this paper is optimal location of distributed generation in electric distribution networks. Load uncertainty and desired voltage range has been modeled using fuzzy data theory. The objective function includes loss reduction, improvement of profile index and voltage stability index with their relevant constraints, voltage constraints and transmittable power from the line. Load variation has been shown for three different time durations (peak, off peak and average).PSO technique has been used to optimize the objective function while Max-Min method has been applied to select the answer. Results produced from the proposed model have been provided in 5 different scenarios on a 33 bus system of IEEE. Manuscript profile

Open Access Article

11 - A New Statistical Characteristics Based Method for Adaptive Learning Rate Adjustment in Learning Automata
M. R. Mollakhalili Meybodi M. R. Meybodi

20.1001.1.16823745.1392.11.2.5.8

The value of learning rate and its change mechanisms is one of the issues in designing learning systems such as learning automata. In most cases a time-based reduction function is used to adjust the learning rate aim at reaching stability in training system. So the lear More

The value of learning rate and its change mechanisms is one of the issues in designing learning systems such as learning automata. In most cases a time-based reduction function is used to adjust the learning rate aim at reaching stability in training system. So the learning rate is a parameter that determines to what extent a learning system is based on past experiences, and the impact of current events on it. This method is efficient but does not properly function in dynamic and non-stationary environments. In this paper, a new method for adaptive learning rate adjustment in learning automata is proposed. In this method, in addition to the length of time to learn, some statistical characteristics of actions probability vector of Learning Automata are used to determine the increase or decrease of learning rate. Furthermore, unlike existing methods, during the process of learning, both increase and decrease of the learning rate is done and Learning Automata responds effectively to changes in the dynamic random environment. Empirical studies show that the proposed method has more flexibility in compatibility to the non-stationary dynamic environments and get out of local maximum points and the learned values are closer to the true values. Manuscript profile

Open Access Article

12 - Left Ventricular Segmentation in Echocardiography Images by Manifold Learning and Dynamic Directed Vector Field Convolution
N. Mashhadi H. Behnam Ahmad Shalbaf Z. Alizadeh Sani

20.1001.1.16823745.1393.12.2.3.3

Cardiac diseases are the major causes of death throughout the world. The study of left ventricular (LV) function is very important in the diagnosis of heart diseases. Automatic tracking of the boundaries of the LV wall during a cardiac cycle is used for quantification o More

Cardiac diseases are the major causes of death throughout the world. The study of left ventricular (LV) function is very important in the diagnosis of heart diseases. Automatic tracking of the boundaries of the LV wall during a cardiac cycle is used for quantification of LV myocardial function in order to diagnose various heart diseases including ischemic disease. In this paper, a new automatic method for segmentation of the LV in echocardiography images of one cardiac cycle by combination of manifold learning and active contour based dynamic directed vector field convolution (DDVFC) is proposed. In this method, first echocardiography images of one cardiac cycle have been embedded in a two dimensional (2-D) space using one of the most popular manifold learning algorithms named Locally Linear Embeddings. In this new space, relationship between these images is well represented. Then, segmentation of the LV wall during a cardiac cycle is done using active contour based DDVFC. In this method, final contour of each segmented frame is used as the initial contour of the next frame. In addition, in order to increase the accuracy of the LV segmentation and also prevent the boundary distortion, maximum range of the active contour motion is limited by Euclidean distances between consequent frames in resultant 2-D manifold. To quantitatively evaluate the proposed method, echoacardiography images of 5 healthy volunteers and 4 patients are used. The results obtained by our method are quantitatively compared to those obtained manually by the highly experienced echocardiographer (gold standard) which depicts the high accuracy of the presented method. Manuscript profile

Open Access Article

13 - Select the Optimal Subset of LABP Features Based on CLA-EC Method in Face Recognition System
A. Hazrati Bishak K. Faez H. Barghi Jond S. Ghatei

20.1001.1.16823745.1393.12.2.8.8

In this paper, we present a new efficient method based on local binary pattern descriptor, for face recognition. Because, the calculations in Local binary pattern are done between two pixels values, so, small changes in the binary pattern affect its performance. In this More

In this paper, we present a new efficient method based on local binary pattern descriptor, for face recognition. Because, the calculations in Local binary pattern are done between two pixels values, so, small changes in the binary pattern affect its performance. In this paper, a new local average binary pattern descriptor is presented based on cellular learning automata and evolutionary computation (CLA-EC). In the proposed method, first, the LABP operator are used to extract uniform local binary patterns from face images; it should be noted that, in LABP operator to obtain more robust feature representation, many sample points has been used. Then, the best subset of patterns found by CLA-EC methods, and the histogram of these patterns is obtained. Finally, support vector machine is used for classification. The results of experiment on FERET data base show the advantage of the proposed algorithm compared to other algorithms. Manuscript profile

Open Access Article

14 - A New eDLA-Based Framework for Finding Optimal Stochastic Sub-Graph
M. R. Mollakhalili Meybodi M. R. Meybodi

20.1001.1.16823745.1393.12.4.1.5

In this paper a new structure of learning automata which is called as extended distributed learning automata (eDLA) is introduced. A new eDLA-based iterative sampling method for finding optimal sub-graph in stochastic graphs is proposed. Some mathematical analysis of th More

In this paper a new structure of learning automata which is called as extended distributed learning automata (eDLA) is introduced. A new eDLA-based iterative sampling method for finding optimal sub-graph in stochastic graphs is proposed. Some mathematical analysis of the proposed algorithm is presented and the convergence property of the algorithm is studied. Our study shows that the proposed algorithm can be converge to the optimal sub-graph. Manuscript profile

Open Access Article

15 - A New Extended Distributed Learning Automata-Based Algorithm for Parameter Learning of a Bayesian Network
M. R. Mollakhalili Meybodi M. R. Meybodi

20.1001.1.16823745.1393.12.4.4.8

In this paper a new learning automata-based algorithm is proposed for learning of parameters of a Bayesian network. For this purpose, a new team of learning automata which is called eDLA is used. In this paper the structure of Bayesian network is assumed to be fixed. Ne More

In this paper a new learning automata-based algorithm is proposed for learning of parameters of a Bayesian network. For this purpose, a new team of learning automata which is called eDLA is used. In this paper the structure of Bayesian network is assumed to be fixed. New arriving sample plays role of the random environment and the accuracy of the current parameters generates the random environment reinforcement signal. Linear algorithm is used to update the action selection probability of the automata. Another key issue in Bayesian networks is parameter learning under circumstances that new samples are incomplete. It is shown that new proposed method can be used in this situation. The experiments show that the accuracy of the proposed automata based algorithm is the same as the traditional enumerative methods such as EM. In addition to the online learning characteristics, the proposed algorithm is in accordance with the conditions in which the data are incomplete and due to the use of learning automaton, has a little computational overhead. Manuscript profile

Open Access Article

16 - Blind Video Steganalysis by Semi-Supervised Approach for Motion Vectors Based Steganography Algorithms
J. Mortazavi Mehrizi M. Khademi H. Sadoghi Yazdi

20.1001.1.16823745.1394.13.4.2.3

Supervised learning algorithms are widely used in blind video steganalysis and the cost of generating labeled data in them is high. That is why only a limited number of steganography algorithms with accessible code can be used for the training the classifier. Therefore, More

Supervised learning algorithms are widely used in blind video steganalysis and the cost of generating labeled data in them is high. That is why only a limited number of steganography algorithms with accessible code can be used for the training the classifier. Therefore, we cannot be sure about the effectiveness of steganalyzer in identifying non-accessible video steganography algorithms. On the other hand, using offline classification methods in the blind video steganalysis causes the learning process be time consuming and the system cannot be updated online. To solve this problem, we propose a new method for the blind video steganalysis by semi-supervised learning approach. In the proposed method, by eliminating the limitation of labeled training dataset, the classifier performance is improved for video steganography algorithms with non-accessible code. It is also proved that the proposed method, compared to common classification methods for the blind video steganalysis, has less time complexity and it is an optimal online technique. The simulation results on the standard database show that in addition to the above advantages, this method has appropriate accuracy and is comparable to common methods. Manuscript profile

Open Access Article

17 - Blind Video Steganalysis by Semi-Supervised Approach for Motion Vectors Based Steganography Algorithms
J. Mortazavi Mehrizi M. Khademi H. Sadoghi Yazdi

Supervised learning algorithms are widely used in blind video steganalysis and the cost of generating labeled data in them is high. That is why only a limited number of steganography algorithms with accessible code can be used for the training the classifier. Therefore, More

Supervised learning algorithms are widely used in blind video steganalysis and the cost of generating labeled data in them is high. That is why only a limited number of steganography algorithms with accessible code can be used for the training the classifier. Therefore, we cannot be sure about the effectiveness of steganalyzer in identifying non-accessible video steganography algorithms. On the other hand, using offline classification methods in the blind video steganalysis causes the learning process be time consuming and the system cannot be updated online. To solve this problem, we propose a new method for the blind video steganalysis by semi-supervised learning approach. In the proposed method, by eliminating the limitation of labeled training dataset, the classifier performance is improved for video steganography algorithms with non-accessible code. It is also proved that the proposed method, compared to common classification methods for the blind video steganalysis, has less time complexity and it is an optimal online technique. The simulation results on the standard database show that in addition to the above advantages, this method has appropriate accuracy and is comparable to common methods. Manuscript profile

Open Access Article

18 - Content Based Image Retrieval by the Fusion of Short Term Learning Methods
B. Bagheri M. Pourmahyabadi H. Nezamabadi-pour

20.1001.1.16823745.1394.13.4.3.4

Content based image retrieval (CBIR) contains a set of techniques to process the visual features of a query image, in order to retrieve images semantically similar to it, in a database. To improve the performance of image retrieval systems, relevance feedback tool can b More

Content based image retrieval (CBIR) contains a set of techniques to process the visual features of a query image, in order to retrieve images semantically similar to it, in a database. To improve the performance of image retrieval systems, relevance feedback tool can be used. In this research, to increase the effectiveness of the image retrieval systems, the fusion of two (multiple) short term learning methods based on relevance feedback is proposed. In the proposed method, fusion is performed in three levels: fusion in ranks, fusion in retrieved images, and fusion in similarities. To evaluate the performance of the proposed method, a CBIR system with 10000 images of 82 different semantic groups is employed. The experimental results confirm the superior of suggested method in terms of retrieval precision. Manuscript profile

Open Access Article

19 - Content Based Image Retrieval by the Fusion of Short Term Learning Methods
B. Bagheri M. Pourmahyabadi H. Nezamabadi-pour

Content based image retrieval (CBIR) contains a set of techniques to process the visual features of a query image, in order to retrieve images semantically similar to it, in a database. To improve the performance of image retrieval systems, relevance feedback tool can b More

Content based image retrieval (CBIR) contains a set of techniques to process the visual features of a query image, in order to retrieve images semantically similar to it, in a database. To improve the performance of image retrieval systems, relevance feedback tool can be used. In this research, to increase the effectiveness of the image retrieval systems, the fusion of two (multiple) short term learning methods based on relevance feedback is proposed. In the proposed method, fusion is performed in three levels: fusion in ranks, fusion in retrieved images, and fusion in similarities. To evaluate the performance of the proposed method, a CBIR system with 10000 images of 82 different semantic groups is employed. The experimental results confirm the superior of suggested method in terms of retrieval precision. Manuscript profile

Open Access Article

20 - Semi-Supervised Learning Based on Extreme Learning
A. Mehrizi H. Sadoghi Yazdi S. J. Seyyed Mahdavi Chabok

20.1001.1.16823745.1394.13.4.4.5

Semi-supervised learning with growing self-organizing map (GSOM) is used in many applications, such as clustering. The main challenges in the Semi-supervised GSOM are calculating parameters such as shape and structure of clustering layer, activation level, and weights o More

Semi-supervised learning with growing self-organizing map (GSOM) is used in many applications, such as clustering. The main challenges in the Semi-supervised GSOM are calculating parameters such as shape and structure of clustering layer, activation level, and weights of classifier layer. Current approaches use initiative methods with a local look have trying to determine these parameters; which its effect, the results of these algorithms is highly dependent on the conditions. This paper studies a semi-supervised learning method based on GSOM and extreme learning for the first time. The proposed method, without the direct calculation of the GSOM parameters and using the extreme learning determines label of each data. Error resulted from the feedback system is used to optimize extreme learning and GSOM. In this paper, in addition to investigating the convergence analysis of the proposed method, sequential extreme learning is also provided for semi-supervised GSOM. Experiments conducted on online and partially labeled data show that the proposed method has a relative advantage in terms of accuracy on semi-supervised GSOM. Manuscript profile

Open Access Article

21 - Learners Grouping in Adaptive Learning Systems Using Fuzzy Grafting Clustering
M. S. Rezaei Gh. A. Montazer

20.1001.1.16823745.1394.13.4.5.6

Quality of adaptive and collaborative learning systems is related to appropriate specifying learners and accuracy of separation learners in homogenous and heterogeneous groups. In the proposed method for learners grouping, researchers effort to improving basic clusterin More

Quality of adaptive and collaborative learning systems is related to appropriate specifying learners and accuracy of separation learners in homogenous and heterogeneous groups. In the proposed method for learners grouping, researchers effort to improving basic clustering methods by combination of them and improving methods. This work makes the complexity of grouping methods increased and quality of result’s groups decreased. In this paper, new method for selection appropriate clusters based on fuzzy theory is proposed. In this method, each cluster is defined as a fuzzy set and the corresponding clusters are determined. So the best cluster is selected among each corresponding clusters. The results of an empirical evaluation of the proposed method based on two criteria: “Davies-Bouldin” and “Purity and Gathering” indicate that this method has better performance than other clustering methods such as FCM, K-means, hybrid clustering method (HCM), evolutionary fuzzy clustering (EFC) and ART neural network. Manuscript profile

Open Access Article

22 - Separating Bichromatic Point Sets by Right Triangles
Z. Moslehi A. Bagheri

20.1001.1.16823745.1394.13.4.9.0

Separating colored point sets is an interesting problem in computational geometry with application in machine learning and pattern recognition. In this problem, we are given a geometric shape C and two point sets P and Q of total size n as red and blue points, respectiv More

Separating colored point sets is an interesting problem in computational geometry with application in machine learning and pattern recognition. In this problem, we are given a geometric shape C and two point sets P and Q of total size n as red and blue points, respectively. Now, we must separate red and blue points by this shape such that all the blue points lie inside it and all the red points lie outside it. In the previous work, we have some algorithms for rectangle and wedge separability but we do not have any algorithm for separating by a triangle and separating by a triangle with a fixed angle such as right triangle. In this paper, we present an efficient algorithm for right triangle seprability. In this algorithm, we use sweep line technique and introduce some events and process them. So, we can report all separating right triangles in O(nlog n) time. Manuscript profile

Open Access Article

23 - Leaning the Structure of Bayesian Networks Using Learning Automata
M. R. Mollakhalili Meybodi M. R. Meybodi

20.1001.1.16823745.1395.14.1.11.3

The structure of a Bayesian network represents a set of conditional independence relations that hold in the domain. Learning the structure of the Bayesian network model that represents a domain can reveal in sights into its underlying causal structure. Automatically lea More

The structure of a Bayesian network represents a set of conditional independence relations that hold in the domain. Learning the structure of the Bayesian network model that represents a domain can reveal in sights into its underlying causal structure. Automatically learning the graph structure of a Bayesian network is a challenge pursued within artificial intelligence studies. In this paper, a new algorithm based on learning automata is proposed for learning the structure of the Bayesian networks. In this algorithm, automata is used as a tool for searching in structure’s space (DAG’s space) of the Bayesian networks. The mathematical behavior of the proposed algorithm is studied. Manuscript profile

Open Access Article

24 - Fusion of Neural Networks Based on Negative Correlation Learning for Offline Handwritten Word Recognition
S. A. A. Abbaszadeh Arani E. Kabir

20.1001.1.16823745.1395.14.1.13.5

In this study, an ensemble classification method, based on negative correlation learning, is used for holistic recognition of handwritten words with limited vocabulary. In this method, training data set, after preprocessing and feature extraction, is applied to the base More

In this study, an ensemble classification method, based on negative correlation learning, is used for holistic recognition of handwritten words with limited vocabulary. In this method, training data set, after preprocessing and feature extraction, is applied to the base Multilayer Perceptron classifiers. These classifiers are trained by negative correlation learning to make them diverse. Features extracted from a test input are applied to the base classifiers, which produce somehow diverse outputs. By combining these outputs, the final output of the system is obtained. For experiments, three feature sets based on zoning, gradient image and contour chain code are extracted from the images. In experiments, performed on 775 images of 31 Province centers from "Iranshahr" dataset, when gradient-based features were used to train 6 Multilayer Perceptron classifiers by negative correlation, by Fusion the outputs of these classifiers through voting, an average recognition rate of 96.10 percent is achieved. Manuscript profile

Open Access Article

25 - Weighted Multi-Level Fuzzy Min-Max Neural Network
R. Davtalab M. A. Balafar M. R. Feizi-Derakhshi

20.1001.1.16823745.1395.14.3.12.8

In this paper a weighted Fuzzy min-max classifier (WL-FMM) which is a type of fuzzy min-max neural network is described. This method is a quick supervised learning tool which capable to learn online and single pass through data. WL-FMM uses smaller size with higher weig More

In this paper a weighted Fuzzy min-max classifier (WL-FMM) which is a type of fuzzy min-max neural network is described. This method is a quick supervised learning tool which capable to learn online and single pass through data. WL-FMM uses smaller size with higher weight to manipulate overlapped area. According to experimental results, proposed method has less time and space complexity rather than other FMM classifiers, and also user manual parameters has less effect on the results of proposed method. Manuscript profile

Open Access Article

26 - An Adaptive Incremental Conductance MPPT Based on BELBIC Controller in Photovoltaic Systems
S. Azimi Sardari B. Mirzaeian Dehkordi M. Niroomand

20.1001.1.16823745.1396.15.2.5.6

Many conventional incremental conductance (INC) methods are applied for maximum power point tracking (MPPT) of photovoltaic arrays. Where, the optimization step size determines the speed of MPPT. Fast tracking could be achieved with bigger increments but the system migh More

Many conventional incremental conductance (INC) methods are applied for maximum power point tracking (MPPT) of photovoltaic arrays. Where, the optimization step size determines the speed of MPPT. Fast tracking could be achieved with bigger increments but the system might not operate properly at the MPP and might become oscillated at this point; therefore, there is a trade-off between the time needed to reach the MPP and the oscillation error. This article is to present an adaptive optimization step size in the INC to improve solar array performance. To adjust the MPP in the photovoltaic (PV) operation point, brain emotional learning based intelligent controller (BELBIC) is applied as an adaptive optimization step size in the INC. This would considerably increase the system's accuracy. The effectiveness of this proposed method is verified by comparing its simulation and experimental results with the conventional methods in different operating conditions. Manuscript profile

Open Access Article

27 - A Cloud-based Learnable Agent-oriented Approach to Control and improve Pacemaker Operation
H. Banki نگار مجمع A. Monadjemi

20.1001.1.16823745.1396.15.4.10.5

This paper aims to present a cloud-based learning agent-oriented approach for verification of the pacemaker behavior by monitoring and heart rate adjustment of an arrhythmic patient. In case of the pacemaker failure or inappropriate heart rate generation, the patient is More

This paper aims to present a cloud-based learning agent-oriented approach for verification of the pacemaker behavior by monitoring and heart rate adjustment of an arrhythmic patient. In case of the pacemaker failure or inappropriate heart rate generation, the patient is put at risk. Using the proposed approach, one can directs the pacemaker rate to correct one when it is incorrect. Using a learnable software agent, the proposed approach is able to learn un-predefined situations and operates accordingly. The proposed approach is cloud based meaning that it sends a message through cloud in case of a critical situation. After determining the patient heart rate by pacemaker, the proposed method verifies this rate against the predefined physician suggestion and automatically corrects it based on a reinforcement learning mechanism if there is some conflict. The proposed method was implemented and installed on a tablet as a patient mobile device for monitoring the pacemaker implanted in the patient chest. The contrast between results of our approach and expected results existing in the dataset showed our approach improved the pacemaker accuracy until 13.24%. The use of the software agent with reinforcement learning is able to play a significant role in improving medical devices in case of critical situations. Manuscript profile

Open Access Article

28 - Geometrical Self-Organizing Map Classifier Based on Active Learning for Steganalysis in the Video Environment by Spending at Least a Label
H. Sadoghi Yazdi A. Mohiaddini M. Khademi

20.1001.1.16823745.1397.16.1.12.8

Classifier is one of the three blocks of a video steganalysis that needs labeled for training. In the blind video steganalysis, due to the lack of access to steganography algorithms, it is difficult to label. In this paper, the semi supervised growing self-organizing ma More

Classifier is one of the three blocks of a video steganalysis that needs labeled for training. In the blind video steganalysis, due to the lack of access to steganography algorithms, it is difficult to label. In this paper, the semi supervised growing self-organizing map classifier has been used to reach the minimum label. For this purpose, a concept called the geometric redundancy of the lower-layer nodes of the semi supervised self-organizing network has been used. It has been shown that this redundancy will create repetitive patterns of the network, so deleting such nodes is possible. Proven due to the existence of one-to-one correspondence between nodes and labels. Reducing nodes leads to a reduction in the number of labels required. The basic point is the need for a geometric redundancy among a number of nodes, which is a conception of abstraction, is the formation of a group by them. Therefore, the proposed algorithm is based on identifying categories and integrating their members. The classifier obtained on this basis has been named a geometric self-organizing map classifier .It is proven that this classifier can achieve the minimum amount of optimal label. The simulation results show a remarkable superiority over the previous algorithms. Manuscript profile

Open Access Article

29 - Adaptive Educational Hypermedia Web Pages Recommending Based on Learning Automata and User Feedback for Resource-Based Learning
Mohammad Tahmasebi Faranak Fotouhi-Dgazvini M. Esmaeili

20.1001.1.16823745.1397.16.1.15.1

Personalized recommender systems and search engines, are two effective key solutions to overcome the information overloading problem. They use the intelligent techniques on users’ interactions to extract their behavioral patterns. These patterns help in making a persona More

Personalized recommender systems and search engines, are two effective key solutions to overcome the information overloading problem. They use the intelligent techniques on users’ interactions to extract their behavioral patterns. These patterns help in making a personalized environment to deliver accurate recommendations. In the technology enhanced learning (TEL) field and in particular resource-based learning, recommendation systems have attracted growing interest. Specially, they are an important module of Adaptive Educational Systems that deliver the appropriate learning objects to their users. Gray-sheep users are a challenge in these systems. They have a little similarity with their peers. So the recommendations to others are not suitable for them. To overcome this problem, we propose the idea of accommodating the user’s learning style to web page features. The user's learning style will be computed according to Felder-Silverman theory. On the other hands, the necessary implicit and explicit meta data will be extracted from OCW web pages. By matching the extracted information, the system delivers the appropriate educational links to user. The user can compare the proposed links, based of our algorithm, to the output of Lucene algorithm. A user’s opinion about every web page as a recommended result would be submitted to the system. The system uses a learning automata algorithm and user’s feedback to deliver best recommendations. The system has been evaluated by a group of engineering students to evaluate its accuracy. Results show that the proposed method outperforms the general search algorithm. This system can be used at formal and informal learning and educational environments for Resource-based learning. Manuscript profile

Open Access Article

30 - Proposing a New Method for Acquiring Skills in Reinforcement Learning with the Help of Graph Clustering
M. Davoodabadi Farahani N. Mozayani

20.1001.1.16823745.1397.16.2.5.3

Reinforcement learning is atype of machine learning methods in which the agent uses its transactions with the environment to recognize the environment and to improve its behavior.One of the main problems of standard reinforcement learning algorithms like Q-learning is t More

Reinforcement learning is atype of machine learning methods in which the agent uses its transactions with the environment to recognize the environment and to improve its behavior.One of the main problems of standard reinforcement learning algorithms like Q-learning is that they are not able to solve large scale problems in a reasonable time. Acquiring skills helps to decompose the problem to a set of sub-problems and to solve it with hierarchical methods. In spite of the promising results of using skills in hierarchical reinforcement learning, it has been shown in some previous studies that based on the imposed task, the effect of skills on learning performance can be quite positive. On the contrary, if they are not properly selected, they can increase the complexity of problem-solving. Hence, one of the weaknesses of previous methods proposed for automatically acquiring skills is the lack of a systematic evaluation method for each acquired skill. In this paper, we propose new methods based on graph clustering for subgoal extraction and acquisition of skills. Also, we present new criteria for evaluating skills, with the help of which, inappropriate skills for solving the problem are eliminated. Using these methods in a number of experimental environments shows a significant increase in learning speed. Manuscript profile

Open Access Article

31 - A Novel Link Prediction Approach on Social Networks
S. Rezavandi Shoaii H. Zare

20.1001.1.16823745.1397.16.3.15.5

Nowadays the network science has been attracted many researchers from a wide variety of different fields and many problems in engineering domains are modelled through social networks measures. One of the most important problems in social networks is the prediction of ev More

Nowadays the network science has been attracted many researchers from a wide variety of different fields and many problems in engineering domains are modelled through social networks measures. One of the most important problems in social networks is the prediction of evolution and structural behavior of the networks that is known as link prediction problem in the related literature. Nowadays people use multiple and different social networks simultaneously and it causes to demonstrate a new domain of research known as heterogenous social networks. There exist a few works on link prediction problem on heterogenous networks. In this paper, first a novel similarity measure for users in heterogenous networks is defined. Then a novel link prediction algorithm is described through a supervised learning approach which is consisted by the generated features from the introduced similarity measures. We employ the standard evaluation criteria for verification of the proposed approach. The comparison of the proposed algorithm to the other well-known earlier works showed that our proposed method has better performance than the other methods based on testing on several network datasets. Manuscript profile

Open Access Article

32 - Semi-Supervised Metric Learning in Stratified Space by Accurate Exploiting of Prior Knowledge
Z. Karimi S. Shiry Ghidary R. Ramezani

20.1001.1.16823745.1397.16.3.17.7

Semi-supervised metric learning has attracted increasing interest in recent years. They enforce smoothness label assumption on the manifold. However, they suffer from two challenges: (1) since data in each class lies on one manifold and the similarity between classes le More

Semi-supervised metric learning has attracted increasing interest in recent years. They enforce smoothness label assumption on the manifold. However, they suffer from two challenges: (1) since data in each class lies on one manifold and the similarity between classes leads the intersection between manifolds, the smoothness assumption on the manifold is violated in intersecting regions. (2) 1NN classifier, which is applied for predicting the label of classes in metric learning methods, is suffered from the rare of labeled data and has not suitable accuracy. In this paper, a novel method for learning semi-supervised metric in the stratified space has been proposed that exploit the prior knowledge, which is the smoothness assumption on each manifold, more accurate than existing methods. In the metric learning stage, it doesn’t apply smoothness assumption on the intersecting regions and in the classification stage, labeled data in the interior regions of manifolds are extended based on the smoothness assumption. The different behavior of the Laplacian of piecewise smooth function on stratified space is exploited for the distinction of the intersecting regions from interior regions of manifolds. The results of experiments verify the improvement of the classification accuracy of the proposed method in the comparison with other methods. Manuscript profile

Open Access Article

33 - Incremental Opinion Mining Using Active Learning over a Stream of Documents
F. Noorbehbahani

20.1001.1.16823745.1397.16.4.12.4

Today, opinion mining is one the most important applications of natural language processing which requires special methods to process documents due to the high volume of comments produced. Since the users’ opinions on social networks and e-commerce websites constitute a More

Today, opinion mining is one the most important applications of natural language processing which requires special methods to process documents due to the high volume of comments produced. Since the users’ opinions on social networks and e-commerce websites constitute an evolving stream, the application of traditional non-incremental classification algorithm for opinion mining leads to the degradation of the classification model as time passes. Moreover, because the users’ comments are massive, it is not possible to label enough comments to build training data for updating the learned model. Another issue in incremental opinion mining is the concept drift that should be supported to handle changing class distributions and evolving vocabulary. In this paper, a new incremental method for polarity detection is proposed which with the application of stream-based active learning selects the best documents to be labeled by experts and updates the classifier. The proposed method is capable of detecting and handling concept drift using a limited labeled data without storing the documents. We compare our method with the state of the art incremental and non-incremental classification methods using credible datasets and standard evaluation measures. The evaluation results show the effectiveness of the proposed method for polarity detection of opinions. Manuscript profile

Open Access Article

34 - Scheduling of IoT Application Tasks in Fog Computing Environment Using Deep Reinforcement Learning
Pegah Gazori Dadmehr Rahbari Mohsen Nickray

20.1001.1.16823745.1399.18.2.9.1

With the advent and development of IoT applications in recent years, the number of smart devices and consequently the volume of data collected by them are rapidly increasing. On the other hand, most of the IoT applications require real-time data analysis and low latency More

With the advent and development of IoT applications in recent years, the number of smart devices and consequently the volume of data collected by them are rapidly increasing. On the other hand, most of the IoT applications require real-time data analysis and low latency in service delivery. Under these circumstances, sending the huge volume of various data to the cloud data centers for processing and analytical purposes is impractical and the fog computing paradigm seems a better choice. Because of limited computational resources in fog nodes, efficient utilization of them is of great importance. In this paper, the scheduling of IoT application tasks in the fog computing paradigm has been considered. The main goal of this study is to reduce the latency of service delivery, in which we have used the deep reinforcement learning approach to meet it. The proposed method of this paper is a combination of the Q-Learning algorithm, deep learning, experience replay, and target network techniques. According to experiment results, The DQLTS algorithm has improved the ASD metric by 76% in comparison to QLTS and 6.5% compared to the RS algorithm. Moreover, it has been reached to faster convergence time than QLTS. Manuscript profile

Open Access Article

35 - Text Generation by a GAN-based Ensemble Approach
Ehsan Montahaie Mahdieh Soleymani Baghshah

20.1001.1.16823745.1399.18.2.10.2

Text generation is one of the important problems in Natural Language Processing field. The former methods for text generation that are based on language modeling by the teacher forcing approach encounter the problem of discrepancy between the training and test phases an More

Text generation is one of the important problems in Natural Language Processing field. The former methods for text generation that are based on language modeling by the teacher forcing approach encounter the problem of discrepancy between the training and test phases and also employing an inappropriate objective (i.e., Maximum Likelihood estimation) for generation. In the past years, Generative Adversarial Networks (GANs) have achieved much popularity due to their capabilities in image generation. These networks have also attracted attention for sequence generation in the last few years. However, since text sequences are discrete, GANs cannot be easily employed for text generation, and new approaches like Reinforcement Learning and approximation have been utilized for this purpose. Furthermore, the instability problem of GANs training causes new challenges. In this paper, a new GAN-based ensemble method is proposed for sequence generation problem. The idea of the proposed method is based on the ratio estimation which enables the model to overcome the problem of discreteness in data. Also, the proposed method is more stable than the other GAN-based methods. It also should be noted that the exposure bias problem of teacher forcing approach does not exist in the proposed method. Experiments show the superiority of the proposed method to previous GAN-based methods for text generation. Manuscript profile

Open Access Article

36 - High Performance Computing via Improvement of Random Forest Algorithm Using Compression and Parallelization Techniques
Naeimeh Mohammad Karimi Mohammad Ghasemzadeh Mahdi Yazdian Dehkordi Amin Nezarat

20.1001.1.16823745.1399.18.2.11.3

This research seeks to promote one of the widely being used algorithms in machine learning, known as the random forest algorithm. For this purpose, we use compression and parallelization techniques. The main challenge we address in this research is about application of More

This research seeks to promote one of the widely being used algorithms in machine learning, known as the random forest algorithm. For this purpose, we use compression and parallelization techniques. The main challenge we address in this research is about application of the random forest algorithm in processing and analyzing big data. In such cases, this algorithm does not show the usual and required performance, due to the needed large number of memory access. This research demonstrates how we can achieve the desired goal by using an innovative compression method, along with parallelization techniques. In this regard, the same components of the trees in the random forest are combined and shared. Also, a vectorization-based parallelization approach, along with a shared-memory-based parallelization method, are used in the processing phase. In order to evaluate its performance, we run it on the Kaggle benchmarks, which are being used widely in machine learning competitions. The experimental results show that contribution of the proposed compression method, could reduce 61% of the required processing time; meanwhile, application of the compression along with the named parallelization methods could lead to about 95% of improvement. Overall, this research implies that the proposed solution can provide an effective step toward high performance computing. Manuscript profile

Open Access Article

37 - Anomaly Detection in the Car Trajectories Using Sparse Reconstruction
Reyhane Taghizade Abbas Ebrahimi moghadam M. Khademi

20.1001.1.16823745.1401.20.1.1.5

In traffic control and vehicle registration systems a big challenge is achieving a system that automatically detects abnormal driving behavior. In this paper a system for detection of vehicle anomalies proposed, which at first extracts spatio-temporal features form clus More

In traffic control and vehicle registration systems a big challenge is achieving a system that automatically detects abnormal driving behavior. In this paper a system for detection of vehicle anomalies proposed, which at first extracts spatio-temporal features form clusters then creates dictionary from these features. This classification stage consists of processes such as, optimized clustering with the bee mating algorithm and sparse processing on spatiotemporal features derived from the training data. Finally the trained classifier is applied to the test data for anomaly detection. The distinction of this study from previous research is using new method of pre-processing to create a dictionary matrix and anomaly detection based on evaluation of matrix that related to each class dependency, which leads to higher accuracy of the proposed method compared to other leading methods. To evaluate the proposed method, UCSD database and video sequences recorded from vehicle traffic on Vakilabad Boulevard at the north side of Ferdowsi University of Mashhad are used and the performance of the proposed method is compare to other competing methods in this field. By analyzing the evaluation standards, we find that the proposed method performance is better than other methods. Manuscript profile

Open Access Article

38 - Bug Detection and Assignment for Mobile Apps via Mining Users' Reviews
Maryam Younesi Abbas Heydarnoori F. Ghanadi

20.1001.1.16823745.1398.17.4.12.1

Increasing the popularity of smart phones and the great ovation of users of mobile apps has turned the app stores to massive software repositories. Therefore, using these repositories can be useful for improving the quality of the program. Since the bridge between users More

Increasing the popularity of smart phones and the great ovation of users of mobile apps has turned the app stores to massive software repositories. Therefore, using these repositories can be useful for improving the quality of the program. Since the bridge between users and developers of mobile apps is the comments that users write in app stores, special attention to these comments from developers can make a dramatic improvement in the final quality of mobile apps. Hence, in recent years, numerous studies have been conducted around the topic of opinion mining, whose intention was to extract and exert important information from user's reviews. One of the shortcomings of these studies is the inability to use the information contained in user comments to expedite and improve the process of fixing the software error. Hence, this paper provides an approach based on users’ feedback for assigning program bugs to developers. This approach builds on the history of a program using its commit data, as well as developers' ability in fixing a program’s errors using the bugs that developers have already resolved in the app. Then, by combining these two criteria, each developer will get a score for her appropriation for considering each review. Next, a list of developers who are appropriate for each bug are provided. The evaluations show that the proposed method would be able to identify the right developer to address the comments with a precision of 74%. Manuscript profile

Open Access Article

39 - A Transfer Learning Algorithm to Improve the Convergence Rate and Accuracy in Cellular Learning Automata
Seyyed Amir Hadi Minoofam Azam Bastanfard M. R. Keyvanpour

20.1001.1.16823745.1400.19.2.11.0

Cellular learning automaton is an intelligent model as a composition of cellular automaton and learning automaton. In this study, an extended algorithm of cellular learning automata is proposed based on transfer learning as the TL-CLA algorithm. In this algorithm, trans More

Cellular learning automaton is an intelligent model as a composition of cellular automaton and learning automaton. In this study, an extended algorithm of cellular learning automata is proposed based on transfer learning as the TL-CLA algorithm. In this algorithm, transfer learning is used as an approach for computation deduction and minimizing the learning cycle. The proposed algorithm is an extended model based on merit function and attitude vector for transfer learning. In the TL-CLA algorithm, the value of the merit function is computed based on the local environment, and the value of the attitude vector is calculated based on the global environment. When these two measures get the threshold values, the transfer of action probabilities causes the transfer learning from the source CLA to the destination CLA. The experimental results show that the proposed TL-CLA model leads to increment the convergence accuracy as 2.7% and 2.2% in two actions and multi-action standard environments, respectively. The improvements in convergence rate are also 8% and 2% in these two environments. The TL-CLA could be applied in knowledge transfer from learning one task to learning another similar task Manuscript profile

Open Access Article

40 - Propose a Proper Algorithm for Incremental Learning Based on Fuzzy Least Square Twin Support Vector Machines
Javad Salimi Sartakhti Salman Goli

20.1001.1.16823745.1400.19.3.3.4

Support Vector machine is one of the most popular and efficient algorithms in machine learning. There are several versions of this algorithm, the latest of which is the fuzzy least squares twin support vector machines. On the other hand, in many machine learning applica More

Support Vector machine is one of the most popular and efficient algorithms in machine learning. There are several versions of this algorithm, the latest of which is the fuzzy least squares twin support vector machines. On the other hand, in many machine learning applications input data is continuously generated, which has made many traditional algorithms inefficient to deal with them. In this paper, for the first time, an incremental version of the fuzzy least squares twin support vector algorithm is presented. The proposed algorithmis represented in both online and quasi-online modes. To evaluate the accuracy and precision of the proposed algorithmfirst we run our algorithm on 6 datasets of the UCI repository. Results showthe proposed algorithm is more efficient than other algorithms (even non-incremental versions). In the second phase in the experiments, we consider an application of Internet of Things, and in particular in data related to daily activities which inherently are incremental. According to experimental results, the proposed algorithm has the best performance compared to other incremental algorithms. Manuscript profile

Open Access Article

41 - A Dynamic Sequential Approach Using Deep Learning to Improve the Performance of Biometrics Match on Card Systems
Mohammad Sabri Mohammad Moin Farbod Razzazi

20.1001.1.16823745.1399.18.1.15.5

Nowadays, the threats such as terrorism and cybercrime are extremely increased, therefore, the identity authentication process is very substantial for the national security of a country. In this paper, we propose a novel multimodal authentication system with sequential More

Nowadays, the threats such as terrorism and cybercrime are extremely increased, therefore, the identity authentication process is very substantial for the national security of a country. In this paper, we propose a novel multimodal authentication system with sequential structure based on deep learning. In the proposed method, feature vectors are extracted automatically through deep network with an end to end architecture. A multi biometric system using two fingerprint and a face is implemented and evaluated. The results demonstrate that, the authentication is done by fingerprints in 91.42% cases and only for 8.58% cases the face modal is required. In addition, the proposed method is more accurate than first and second fingerprint by 35% and 30% at FMR=0.001, respectively. As a result, we augmented the accuracy of the system and at the same time reduced the acquisition and matching time. This conducts to the improvement of user convenience and security of the service provider, simultaneously. The achievements of this work can be used to increase the effectiveness of authentication process and can play an important role in the acceptability of real world applications. Manuscript profile

Open Access Article

42 - Spatio-Temporal Prediction of Vegetation Dynamics Based on Remote Sensing Data Using Deep Learning
Elham Zangeneh H. Mashayekhi Saeed Gharachelo

20.1001.1.16823745.1401.20.4.2.2

Understanding and analyzing spatial-temporal data changes is very important in various applications, including the protection and development of natural resources. In the past studies, Markov process and comparison-based methods were mainly used to predict the changes o More

Understanding and analyzing spatial-temporal data changes is very important in various applications, including the protection and development of natural resources. In the past studies, Markov process and comparison-based methods were mainly used to predict the changes of vegetation indices, whose accuracy still needs improvement. Although time series analysis has been used to predict some indices, the method to extract these indices from remote sensing data and model their sequences with deep learning is rarely observed. In this article, a method for predicting changes in plant indices based on deep learning is presented. The research data includes Landsat satellite images from 2000 to 2018, related to four seasons in the north and east of Shahrood city in Semnan province. The time span of the extracted images makes it possible to predict changes in vegetation cover. Vegetation indices extracted from the data set are NDVI, SAVI and RVI. After performing atmospheric corrections on the images, the desired indicators are extracted and then the data is converted into a time series. Finally, the modeling of the sequence of these data is performed by the Short-Long-Term Memory (LSTM) network. The results of the experiments show that the deep network is able to predict future values with high accuracy. The amount of the model error without additional data is 0.03 for the NDVI index, 0.02 for the SAVI index, and 0.06 for the RVI index. Manuscript profile

Open Access Article

43 - Detecting Human Activities Based on Motion Sensors in IOT Using Deep Learning
Abbas Mirzaei fatemeh faraji

20.1001.1.16823745.1400.19.4.10.3

Control of areas and locations and motion sensors in the Internet of Things requires continuous control to detect human activities in different situations, which is an important challenge, including manpower and human error. Permanent human control of IoT motion sensors More

Control of areas and locations and motion sensors in the Internet of Things requires continuous control to detect human activities in different situations, which is an important challenge, including manpower and human error. Permanent human control of IoT motion sensors also seems impossible. The IoT is more than just a simple connection between devices and systems. IoT information sensors and systems help companies get a better view of system performance. This study presents a method based on deep learning and a 30-layer DNN neural network for detecting human activity on the Fordham University Activity Diagnostic Data Set. The data set contains more than 1 million lines in six classes to detect IoT activity. The proposed model had almost 90% and an error rate of 0.22 in the evaluation criteria, which indicates the good performance of deep learning in activity recognition. Manuscript profile

Open Access Article

44 - A New Algorithm Based on Distributed Learning Automata for Solving Stochastic Linear Optimization Problems on the Group of Permutations
mohammadreza mollakhalili meybodi masoumeh zojaji

20.1001.1.16823745.1401.20.2.2.8

In the present research, a type of permutation optimization was introduced. It is assumed that the cost function has an unknown probability distribution function. Since the solution space is inherently large, solving the problem of finding the optimal permutation is com More

In the present research, a type of permutation optimization was introduced. It is assumed that the cost function has an unknown probability distribution function. Since the solution space is inherently large, solving the problem of finding the optimal permutation is complex and this assumption increases the complexity. In the present study, an algorithm based on distributed learning automata was presented to solve the problem by searching in the permutation answer space and sampling random values. In the present research, in addition to the mathematical analysis of the behavior of the proposed new algorithm, it was shown that by choosing the appropriate values of the parameters of the learning algorithm, this new method can find the optimal solution with a probability close to 100% and by targeting the search using the distributed learning algorithms. The result of adopting this policy is to decrease the number of samplings in the new method compared to methods based on standard sampling. In the following, the problem of finding the minimum spanning tree in the stochastic graph was evaluated as a random permutation optimization problem and the proposed solution based on learning automata was used to solve it. Manuscript profile

Open Access Article

45 - An Intelligent Vision System for Automatic Forest Fire Surveillance
Mohammad Sadegh Kayhanpanah Behrooz Koohestani

20.1001.1.16823745.1401.20.1.3.7

Fighting forest fires to avoid their potential dangers as well as protect natural resources is a challenge for researchers. The goal of this research is to identify the features of fire and smoke from the unmanned aerial vehicle (UAV) visual images for classification, o More

Fighting forest fires to avoid their potential dangers as well as protect natural resources is a challenge for researchers. The goal of this research is to identify the features of fire and smoke from the unmanned aerial vehicle (UAV) visual images for classification, object detection, and image segmentation. Because forests are highly complex and nonstructured environments, the use of the vision system is still having problems such as the analogues of flame characteristics to sunlight, plants, and animals, or the smoke blocking the images of the fire, which causes false alarms. The proposed method in this research is the use of convolutional neural networks (CNNs) as a deep learning method that can automatically extract or generate features in different layers. First, we collect data and increase them according to data augmentation methods, and then, the use of a 12-layer network for classification as well as transfer learning method for segmentation of images is proposed. The results show that the data augmentation method used due to resizing and processing the input images to the network to prevent the drastic reduction of the features in the original images and also the CNNs used can extract the fire and smoke features in the images well and finally detect and localize them. Manuscript profile

Open Access Article

46 - Efficient Recognition of Human Actions by Limiting the Search Space in Deep Learning Methods
m. koohzadi N. Moghadam

20.1001.1.16823745.1400.19.3.6.7

The efficiency of human action recognition systems depends on extracting appropriate representations from the video data. In recent years, deep learning methods have been proposed to extract efficient spatial-temporal representations. Deep learning methods, on the other More

The efficiency of human action recognition systems depends on extracting appropriate representations from the video data. In recent years, deep learning methods have been proposed to extract efficient spatial-temporal representations. Deep learning methods, on the other hand, have a high computational complexity for development over temporal domain. Challenges such as the sparsity and limitation of discriminative data, and highly noise factors increase the computational complexity of representing human actions. Therefore, creating a high accurate representation requires a very high computational cost. In this paper, spatial and temporal deep learning networks have been enhanced by adding appropriate feature selection mechanisms to reduce the search space. In this regard, non-online and online feature selection mechanisms have been studied to identify human actions with less computational complexity and higher accuracy. The results showed that the non-linear feature selection mechanism leads to a significant reduction in computational complexity and the online feature selection mechanism increases the accuracy while controlling the computational complexity. Manuscript profile

Open Access Article

47 - Deep Extreme Learning Machine: A Combined Incremental Learning Approach for Data Stream Classification
Javad Hamidzadeh Mona Moradi

20.1001.1.16823745.1401.20.1.5.9

Streaming data refers to data that is continuously generated in the form of fast streams with high volumes. This kind of data often runs into evolving environments where a change may affect the data distribution. Because of a wide range of real-world applications of dat More

Streaming data refers to data that is continuously generated in the form of fast streams with high volumes. This kind of data often runs into evolving environments where a change may affect the data distribution. Because of a wide range of real-world applications of data streams, performance improvement of streaming analytics has become a hot topic for researchers. The proposed method integrates online ensemble learning into extreme machine learning to improve the data stream classification performance. The proposed incremental method does not need to access the samples of previous blocks. Also, regarding the AdaBoost approach, it can react to concept drift by the component weighting mechanism and component update mechanism. The proposed method can adapt to the changes, and its performance is leveraged to retain high-accurate classifiers. The experiments have been done on benchmark datasets. The proposed method can achieve 0.90% average specificity, 0.69% average sensitivity, and 0.87% average accuracy, indicating its superiority compared to two competing methods. Manuscript profile

Open Access Article

48 - A Prediction-Based Load Distribution Approach for Software-Defined Networks
Hossein Mohammadi سیداکبر مصطفوی

20.1001.1.16823745.1400.19.4.15.8

Software-defined networking is a new network architecture which separates the control layer from the data layer. In this approach, the responsibility of the control layer is delegated to the controller software to dynamically determine the behavior of the entire network More

Software-defined networking is a new network architecture which separates the control layer from the data layer. In this approach, the responsibility of the control layer is delegated to the controller software to dynamically determine the behavior of the entire network. It results in a flexible network with centralized management in which network parameters can be well controlled. Due to the increasing number of users, the emergence of new technologies, the explosive growth of network traffic, meeting the requirements of quality of service and preventing underload or overload of resources, load balancing in software-based networks is of substantial importance. Load imbalance increases costs, reduces scalability, flexibility, efficiency, and delay in network service. So far, a number of solutions have been proposed to improve the performance and load balancing in the network, which take into account different criteria such as power consumption and server response time, but most of them do not prevent the system from entering the load imbalance mode and the risks of load imbalance. In this paper, a predictive load balancing method is proposed to prevent the system from entering the load imbalance mode using the Extreme Learning Machine (ELM) algorithm. The evaluation results of the proposed method show that in terms of controller processing delay, load balance and response time, it performs better than CDAA and PSOAP methods. Manuscript profile

Open Access Article

49 - Iranian Dastgah Music Recognition Based on Notes Sequence Extraction and Use of LSTM Networks
سینا غضنفری پور M. Khademi Abbas Ebrahimi moghadam

20.1001.1.16823745.1401.20.2.6.2

Iranian "Dastgah" music classification by computer is a very interesting yet complex and challenging topic for those who are interested in Iranian Dastgah music. The aforementioned problem is important, firstly, due to its many applications in different areas such as co More

Iranian "Dastgah" music classification by computer is a very interesting yet complex and challenging topic for those who are interested in Iranian Dastgah music. The aforementioned problem is important, firstly, due to its many applications in different areas such as composing and teaching music, and secondly, because of the needs of ordinary people to computer to detect the Dastgah. This paper presents a method for recognition of the genre (Dastgah) and subgenre (sub-Dastgah) of Iranian music based on sequential note extraction, hierarchical classification, and the use of LSTM networks. In the proposed method, the music track is first classified into one of the three general categories. The first category includes only "Mahour" Dastgah, the second category includes "Shour" and "Nava", and the third category includes "Homayoun", "Segah" and "Chahargah". Then, for each category, depending on its type, a different number of classifiers are applied until one of the 6 Dastgah and 11 sub-Dastgah of Iranian music are recognized. This research is not limited to any particular style of playing or instruments, it is also not affected by neither the speed nor the techniques of player. The labeled tracks in the "Arg" database, which is created for this research, are solo. However, some of them are also played by percussion instruments (such as the Tombak) along with melodic instruments. The results show that recognition of 6 main Dastgah and 11 sub-Dastgah have been approved by an average accuracy of 74.5% and 66.35%, respectively, which is more promising compared to other few similar studies. Manuscript profile

Open Access Article

50 - Load Balancing in Fog Nodes using Reinforcement Learning Algorithm
niloofar tahmasebi pouya Mehdi-Agha Sarram

20.1001.1.16823745.1401.20.4.3.3

Fog computing is an emerging research field for providing cloud computing services to the edges of the network. Fog nodes process data stream and user requests in real-time. In order to optimize resource efficiency and response time, increase speed and performance, task More

Fog computing is an emerging research field for providing cloud computing services to the edges of the network. Fog nodes process data stream and user requests in real-time. In order to optimize resource efficiency and response time, increase speed and performance, tasks must be evenly distributed among the fog nodes. Therefore, in this paper, a new method is proposed to improve the load balancing in the fog computing environment. In the proposed algorithm, when a task is sent to the fog node via mobile devices, the fog node using reinforcement learning decides to process that task itself, or assign it to one of the neighbor fog nodes or cloud for processing. The evaluation shows that the proposed algorithm, with proper distribution of tasks between nodes, has less delay to tasks processing than other compared methods. Manuscript profile

Open Access Article

51 - Generation of Persian sentences By Generative Adversarial Network
Nooshin riahi Sahar Jandaghy

20.1001.1.16823745.1402.21.1.10.1

Text generation is a field of natural language processing. Text generation enables the system to produce comprehensive, .grammatically correct texts like humans. Applications of text generation include image Captioning, poetry production, production of meteorological re More

Text generation is a field of natural language processing. Text generation enables the system to produce comprehensive, .grammatically correct texts like humans. Applications of text generation include image Captioning, poetry production, production of meteorological reports and environmental reports, production of business reports, automatic text summarization, .With the appearance of deep neural networks, research in the field of text generation has change to use of these networks, but the most important challenge in the field of text generation using deep neural networks is the data is discrete, which has made gradient inability to transmit. Recently, the use of a new approach in the field of deep learning, called generative adversarial networks (GANs) for the generation of image, sound and text has been considered. The purpose of this research is to use this approach to generate Persian sentences. In this paper, three different algorithms of generative adversarial networks were used to generate Persian sentences. to evaluate our proposed methods we use BLEU and self-BLEU because They compare the sentences in terms of quality and variety. Manuscript profile

Open Access Article

52 - Provide a Personalized Session-Based Recommender System with Self-Attention Networks
Azam Ramazani A. Zareh

20.1001.1.16823745.1401.20.3.3.1

Session-based recommender systems predict the next behavior or interest of the user based on user behavior and interactions in a session, and suggest appropriate items to the user accordingly. Recent studies to make recommendations have focused mainly on the information More

Session-based recommender systems predict the next behavior or interest of the user based on user behavior and interactions in a session, and suggest appropriate items to the user accordingly. Recent studies to make recommendations have focused mainly on the information of the current session and ignore the information of the user's previous sessions. In this paper, a personalized session-based recommender model with self-attention networks is proposed, which uses the user's previous recent sessions in addition to the current session. The proposed model uses self-attention networks (SANs) to learn the global dependencies among all session items. First, SAN is trained based on anonymous sessions. Then for each user, the sequences of the current session and previous sessions are given to the network separately, and by weighted combining the ranking results from each session, the final recommended items are obtained. The proposed model is tested and evaluated on real-world Reddit dataset in two criteria of accuracy and mean reciprocal rank. Comparing the results of the proposed model with previous approaches indicates the ability and effectiveness of the proposed model in providing more accurate recommendations. Manuscript profile

Open Access Article

53 - A Content-Based Image Retrieval System Using Semi-Supervised Learning and Frequent Patterns Mining
Maral Kolahkaj

20.1001.1.16823745.1401.20.3.7.5

Content-based image retrieval, which is also known as query based on image content, is one of the sub-branches of machine vision, which is used to organize and recognize the content of digital images using visual features. This technology automatically searches the imag More

Content-based image retrieval, which is also known as query based on image content, is one of the sub-branches of machine vision, which is used to organize and recognize the content of digital images using visual features. This technology automatically searches the images similar to the query image from huge image database and it provides the most similar images to the users by directly extracting visual features from image data; not keywords and textual annotations. Therefore, in this paper, a method is proposed that utilizes wavelet transformation and combining features with color histogram to reduce the semantic gap between low-level visual features and high-level meanings of images. In this regard, the final output will be presented using the feature extraction method from the input images. In the next step, when the query images are given to the system by the target user, the most similar images are retrieved by using semi-supervised learning that results from the combination of clustering and classification based on frequent patterns mining. The experimental results show that the proposed system has provided the highest level of effectiveness compared to other methods. Manuscript profile

Open Access Article

54 - SQ-PUF: A Resistant PUF-Based Authentication Protocol against Machine-Learning Attack
Abolfazl Sajadi Bijan Alizadeh

Physically unclonable functions (PUFs) provide hardware to generate a unique challenge-response pattern for authentication and encryption purposes. An essential feature of these circuits is their unpredictability, meaning that an adversary cannot sufficiently predict fu More

Physically unclonable functions (PUFs) provide hardware to generate a unique challenge-response pattern for authentication and encryption purposes. An essential feature of these circuits is their unpredictability, meaning that an adversary cannot sufficiently predict future responses from previous observations. However, machine learning algorithms have been demonstrated to be a severe threat to PUFs since they are capable of accurately modeling their behavior. In this work, we analyze PUF security threats and propose a PUF-based authentication mechanism called SQ-PUF, which can provide good resistance to machine learning attacks. In order to make it harder to simulate or predict, we obfuscated the correlation between challenge-response pairs. Experimental results show that, unlike existing PUFs, even with a large data set, the SQ-PUF model cannot be successfully attacked with a maximum prediction accuracy of 53%, indicating that this model is unpredictable. In addition, the uniformity in this model remains almost the same as the ideal value in A-PUF. Manuscript profile

Open Access Article

55 - Stock Price Movement Prediction Using Directed Graph Attention Network
Alireza Jafari Saman Haratizadeh

20.1001.1.16823745.1402.21.1.13.4

Prediction of the future behavior of the stock market has always attracted researchers' attention as an important challenge in the field of machine learning. In recent years deep learning methods have been successfully applied in this domain to improve prediction perfor More

Prediction of the future behavior of the stock market has always attracted researchers' attention as an important challenge in the field of machine learning. In recent years deep learning methods have been successfully applied in this domain to improve prediction performance. Previous studies have demonstrated that aggregating information from related stocks can improve the performance of prediction. However, the capacity of modeling the stocks relations as directed graphs and the power of sophisticated graph embedding techniques such as Graph Attention Networks have not been exploited so far for prediction in this domain. In this work, we introduce a framework called DeepNet that creates a directed graph representing how useful the data from each stock can be for improving the prediction accuracy of any other stocks. DeepNet then applies Graph Attention Network to extract a useful representation for each node by aggregating information from its neighbors, while the optimal amount of each neighbor's contribution is learned during the training phase. We have developed a novel Graph Attention Network model called DGAT that is able to define unequal contribution values for each pair of adjacent nodes in a directed graph. Our evaluation experiments on the Tehran Stock Exchange data show that the introduced prediction model outperforms the state-of-the-art baseline algorithms in terms of accuracy and MCC measures. Manuscript profile

Open Access Article

56 - Social Networks Embedding Based on the Employment of Community Recognition and Latent Semantic Feature Extraction Approaches
Mohadeseh Taherparvar Fateme Ahmadi abkenari Peyman bayat

The purpose of embedding social networks, which has recently attracted a lot of attention, is to learn to display in small dimensions for each node in the network while maintaining the structure and characteristics of the network. In this paper, we propose the effect of More

The purpose of embedding social networks, which has recently attracted a lot of attention, is to learn to display in small dimensions for each node in the network while maintaining the structure and characteristics of the network. In this paper, we propose the effect of identifying communities in different situations such as community detection during or before the process of random walking and also the effect of semantic textual information of each node on network embedding. Then two main frameworks have been proposed with community and context aware network embedding and community and semantic feature-oriented network embedding. In this paper, in community and context aware network embedding, the detection of communities before the random walk process, is performed through using the EdMot non-overlapping method and EgoNetSplitter overlapping method. However, in community and semantic feature-oriented network embedding, the recognition of communities during a random walk event is conducted using a Biterm topic model. In all the proposed methods, text analysis is examined and finally, the final display is performed using the Skip-Gram model in the network. Experiments have shown that the methods proposed in this paper work better than the superior network embedding methods such as Deepwalk, CARE, CONE, and COANE and have reached an accuracy of nearly 0.9 and better than other methods in terms of edge prediction criteria in the network. Manuscript profile

Open Access Article

57 - A Fast and Lightweight Network for Road Lines Detection Using Mobile-Net Architecture and Different Loss Functions
Pejman Goudarzi milad Heydari Mehdi Hosseinpour

By using the line detection system, the relative position of the self-driving cars compared to other cars, the possibility of leaving the lane or an accident can be checked. In this paper, a fast and lightweight line detection approach for images taken from a camera ins More

By using the line detection system, the relative position of the self-driving cars compared to other cars, the possibility of leaving the lane or an accident can be checked. In this paper, a fast and lightweight line detection approach for images taken from a camera installed in the windshield of cars is presented. Most of the existing methods consider the problem of line detection in the form of classification at the pixel level. These methods despite having high accuracy, suffer from two weaknesses of having the high computational cost and not paying attention to the general lines content information of the image (as a result, they cannot detect if there is an obstacle). The proposed method checks the presence of lines in each row by using the row-based selection method. Also, the use of Mobile-net architecture has led to good results with fewer learning parameters. The use of three different functions as cost functions, with different objectives, has resulted in obtaining excellent results and considering general content information along with local information. Experiments conducted on the TuSimple video image collection show the suitable performance of the proposed approach both in terms of efficiency and especially in terms of speed. Manuscript profile

Open Access Article

58 - Semantic Word Embedding Using BERT on the Persian Web
shekoofe bostan Ali-Mohammad Zare-Bidoki mohamad reza pajohan

20.1001.1.16823745.1402.21.2.11.4

Using the context and order of words in sentence can lead to its better understanding and comprehension. Pre-trained language models have recently achieved great success in natural language processing. Among these models, The BERT algorithm has been increasingly popular More

Using the context and order of words in sentence can lead to its better understanding and comprehension. Pre-trained language models have recently achieved great success in natural language processing. Among these models, The BERT algorithm has been increasingly popular. This problem has not been investigated in Persian language and considered as a challenge in Persian web domain. In this article, the embedding of Persian words forming a sentence was investigated using the BERT algorithm. In the proposed approach, a model was trained based on the Persian web dataset, and the final model was produced with two stages of fine-tuning the model with different architectures. Finally, the features of the model were extracted and evaluated in document ranking. The results obtained from this model are improved compared to results obtained from other investigated models in terms of accuracy compared to the multilingual BERT model by at least one percent. Also, applying the fine-tuning process with our proposed structure on other existing models has resulted in the improvement of the model and embedding accuracy after each fine-tuning process. This process will improve result in around 5% accuracy of the Persian web ranking. Manuscript profile

Open Access Article

59 - Machine Learning-Based Security Resource Allocation for Defending against Attacks in the Internet of Things
Nasim Navaei Vesal Hakami

Nowadays, the Internet of Things (IoT) has become the focus of security attacks due to the limitation of processing resources, heterogeneity, energy limitation in objects, and the lack of a single standard for implementing security mechanisms. In this article, a solutio More

Nowadays, the Internet of Things (IoT) has become the focus of security attacks due to the limitation of processing resources, heterogeneity, energy limitation in objects, and the lack of a single standard for implementing security mechanisms. In this article, a solution will be presented for the problem of security resources allocating to deal with attacks in the Internet of Things. Security Resource Allocation (SRA) problem in the IoT networks refers to the placement of the security resources in the IoT infrastructure. To solve this problem, it is mandatory to consider the dynamic nature of the communication environments and the uncertainty of the attackers' actions. In the traditional approaches for solving the SRA, the attacker works over based on his assumptions about the system conditions. Meanwhile, the defender collects the system's information with prior knowledge of the attacker's behavior and the targeted nodes. Unlike the mentioned traditional approaches, this research has adopted a realistic approach for the Dynamic Security Resources Allocation in the IoT to battle attackers with unknown behavior. In the stated problem, since there is a need to decide on deploying several security resources during the learning periods, the state space of the strategies is expressed in the combinatorial form. Also, the SRAIoT problem is defined as a combinatorial-adversarial multi-armed bandit problem. Since switching in the security resources has a high cost, in real scenarios, this cost is included in the utility function of the problem. Thus, the proposed framework considers the switching cost and the earned reward. The simulation results show a faster convergence of the weak regret criterion of the proposed algorithms than the basic combinatorial algorithm. In addition, in order to simulate the IoT network in a realistic context, the attack scenario has been simulated using the Cooja simulator. Manuscript profile

Open Access Article

60 - Identification of Cancer-Causing Genes in Gene Network Using Feedforward Neural Network Architecture
مصطفی اخوان صفار abbas ali rezaee

Identifying the genes that initiate cancer or the cause of cancer is one of the important research topics in the field of oncology and bioinformatics. After the mutation occurs in the cancer-causing genes, they transfer it to other genes through protein-protein interact More

Identifying the genes that initiate cancer or the cause of cancer is one of the important research topics in the field of oncology and bioinformatics. After the mutation occurs in the cancer-causing genes, they transfer it to other genes through protein-protein interactions, and in this way, they cause cell dysfunction and the occurrence of disease and cancer. So far, various methods have been proposed to predict and classify cancer-causing genes. These methods mostly rely on genomic and transcriptomic data. Therefore, they have a low harmonic mean in the results. Research in this field continues to improve the accuracy of the results. Therefore, network-based methods and bioinformatics have come to the aid of this field. In this study, we proposed an approach that does not rely on mutation data and uses network methods for feature extraction and feedforward three-layer neural network for gene classification. For this purpose, the breast cancer transcriptional regulatory network was first constructed. Then, the different features of each gene were extracted as vectors. Finally, the obtained vectors were given to a feedforward neural network for classification. The obtained results show that the use of methods based on multilayer neural networks can improve the accuracy and harmonic mean and improve the performance compared to other computational methods. Manuscript profile

Open Access Article

61 - Design and Collection of Speech Data as the First Step of Localization the Intelligent Diagnosis of Autism in Iranian Children
Maryam Alizadeh Shima tabibian

Autism Spectrum Disorder is a type of disorder in which, the patients suffer from a developmental disorder that manifests itself by symptoms such as inability to social communication. Thus, the most apparent sign of autism is a speech disorder. The first part of this pa More

Autism Spectrum Disorder is a type of disorder in which, the patients suffer from a developmental disorder that manifests itself by symptoms such as inability to social communication. Thus, the most apparent sign of autism is a speech disorder. The first part of this paper reviews research studies conducted to automatically diagnose autism based on speech processing methods. According to our review, the main speech processing approaches for diagnosing autism can be divided into two groups. The first group detects individuals with autism by processing their answers or feelings in response to questions or stories. The second group distinguishes people with autism from healthy people because of the accuracy of recognizing their spoken utterances based on automatic speech recognition systems. Despite much research being conducted outside Iran, few studies have been conducted in Iran. The most important reason for this is the lack of rich data that meet the needs of autism diagnosis based on the speech processing of suspected people. In the second part of the paper, we discuss the process of designing, collecting, and evaluating a speaker-independent dataset for autism diagnosis in Iranian children as the first step in the localization of the mentioned field. Manuscript profile

Open Access Article

62 - Combination of Instance Selection and Data Augmentation Techniques for Imbalanced Data Classification
Parastoo Mohaghegh Samira Noferesti Mehri Rajaei

Mohaghegh, S. Noferesti*, and M. Rajaei Abstract: In the era of big data, automatic data analysis techniques such as data mining have been widely used for decision-making and have become very effective. Among data mining techniques, classification is a common method fo More

Mohaghegh, S. Noferesti*, and M. Rajaei Abstract: In the era of big data, automatic data analysis techniques such as data mining have been widely used for decision-making and have become very effective. Among data mining techniques, classification is a common method for decision making and prediction. Classification algorithms usually work well on balanced datasets. However, one of the challenges of the classification algorithms is how to correctly predicting the label of new samples based on learning on imbalanced datasets. In this type of dataset, the heterogeneous distribution of the data in different classes causes examples of the minority class to be ignored in the learning process, while this class is more important in some prediction problems. To deal with this issue, in this paper, an efficient method for balancing the imbalanced dataset is presented, which improves the accuracy of the machine learning algorithms to correct prediction of the class label of new samples. According to the evaluations, the proposed method has a better performance compared to other methods based on two common criteria in evaluating the classification of imbalanced datasets, namely "Balanced Accuracy" and "Specificity". Manuscript profile

Current Issue

Published Issues

Menu

Browse

List of Articles learning