فصلنامه پردازش علائم و داده ها
سال بیستم شماره 1 (پیاپی 55، بهار 1402)

The growth of vehicles and the increase in urban traffic have led to the need for safe, secure and efficient Intelligent Transportation Systems (ITS). Improving the efficiency of ITS needs an efficient communication system. To this end, vehicular networks have been considered as a communication infrastructure in ITS. One of the mechanisms to manage vehicles in ITS is to categorize vehicles into platoon. Platoon is proposed as a way to increase road capacity, improve safety, minimize travel time, increase fuel efficiency, reduce environmental impact, lower-traffic jams and facilitate driving. To create a platoon through autonomous cooperative driving, vehicles must be able to communicate wirelessly, which is possible through vehicular networks. Wireless communications in vehicular networks suffer from three major problems, which are: limited range of radiofrequency, high data transmission over the wireless network, as well as security problems. Using omnidirectional antennas and utilizing DSRC channels in the lower layers are the base solutions to overcome these issues but, Omnidirectional antennas spread the vulnerability to all nodes within the signal propagation range and congestion challenges in the DSRC channel intensify the collision of packets. Such issues decrease the platoon's stability and security. Recent studies on vehicular networks security have focused on presenting solutions to reduce vulnerabilities at the level of communications. However, while the security of communications has been extensively explored in previous work, the security of platoons has recently been considered by researchers. In vehicular networks, an attack can pose a threat to the security and privacy of the platoon, so the security of the platoon has received a great deal of attention.  To improve the security of the platoon, the SP-VLC solution is proposed. This method uses asymmetric encryption and the transmission of information through Visible Light Communications (VLC) to overcome security challenges. In intra-platoon attacks, the attacker is a member of the platoon and has access to all keys and encrypted information, so this solution is not effective against platoon’s internal attacks. Moreover, the SP-VLC solution uses only the data flow topology of predecessor-following to transfer data. Another solution which fixes internal attack problem of SP-VLC solution is voting method. But this solution does not work well for the "predecessor-following" information flow topology because vehicles in the voting process need to receive information from at least two vehicles, while in the mentioned topology, vehicles receive information only from the vehicle in front.   Therefore, the voting solution does not have the required efficiency in "predecessor-following" topology. To overcome the challenges of the voting method, a new solution has been proposed that takes advantage of changes in the distance of each vehicle from the leader vehicle and the best speed of the malicious vehicle compared to the leader speed before the attack. Also, in order to evaluate the efficiency of the proposed solution, evaluation in Bidirectional-leader (BDL) and Predecessor-leader following (PLF) topologies has been performed as a representative of other information flow topologies. The performance of the proposed solution has been evaluated using OMNeT ++, SUMO, and PLEXE tools. To measure the efficiency of the proposed method, the parameters of position error, speed error and control effort have been used. The simulation of the proposed solution shows that this study responds to spoofing attack in 0.43 seconds and to message falsification attack in 0.48 seconds.

Fog computing address numerous cloud computing challenges such as high latency, low capacity and network failure.  The cloud computing infrastructure includes a large number of IoT devices with the ability to process in the cloud environment.  In fog computing, processing and storage provide on IoT devices locally instead of remote servers, therefore, fog computing is the best choice to enable IoT in order to provide efficient, faster and secure services for many users on the edge of the network.  Fog computing have a variety of challenges. One of these important challenges is resource management and task scheduling such that solving this problem has a great impact on system efficiency and service quality. In this paper, we present a task scheduling approach based on the imperialist competition algorithm namely, Imperialist Competitive Algorithm-based Task Scheduling in Fog Computing (ICTF). In the proposed method, we consider the search space as a directional graph. Assume that each task that contains a set of tasks is a graph with a root node and an end leaf node whose middle nodes are the task set. Each path in this graph that starts at the root node and ends at the leaf is a solution represented by a string. This solution is modeled as a country. Therefore, in the proposed method, the concept of country includes the tasks of a job along with the fog nodes that are assigned to these tasks. The initial population consists of a random number of these solutions. ICTF presents a cost function consisting of three important criteria for assessing the initial population of countries and determining the imperialists and colonies countries include energy, execution time and execution cost. The assimilation operation is performed on two different members of countries, namely the imperialists and colonies country, and two new types of members are created called children. The countries participating in this process are among the best countries and are selected using the cost function. In this process, the best offspring produced are passed on to the next generation, and this operation continues until the final population of the countries is obtained. The assimilation operator has different models and in this article we use the two-point assimilation operator. The name of the revolution operator used in this algorithm is the inverse of the task. This operator randomly selects two tasks belonging to a fog node and moves them together. The above operation is repeated until the population converges and reaches the final answer. We show that our proposed approach is more efficient in terms of makespan, resource utilization, energy consumption and remaining energy compared to the similar approach.

Due to the increasing information and the detailed analysis of them, the clustering problems that detect the hidden patterns lie in the data are still of great importance. On the other hand, clustering of high-dimensional data using previous traditional methods has many limitations. In this study, a semi-supervised ensemble clustering method is proposed for a set of high-dimensional medical data. In the proposed method of this study, little information is available as prior knowledge using the information on similarity or dissimilarity (as a number of pairwise constraints). Initially using the transitive property, we generalize the pairwise constraints to all data. Then we divide the feature space into a number of sub-spaces, and to find the optimal clustering solution, the feature space is divided into an unequal number of sub-spaces randomly. A semi-supervised spectral clustering based on the p-Laplacian graph is performed at each sub-space independently. Specifically, to increase the accuracy of spectral clustering, we have used the spectral clustering method based on the p-Laplacian graph. The p-Laplacian graph is a nonlinear generalization of the Laplacian graph. The results of any clustering solutions are compared with the pairwise constraints and according to the level of matching, a degree of confidence is assigned to each clustering solution. Based on these degrees of confidence, an ensemble adjacency matrix is formed, which is the result of considering the results of all clustering solutions for each sub-space. This ensemble adjacency matrix is used in the final spectral clustering algorithm to find the clustering solution of the whole sub-space. Since the sub-spaces are generated randomly with an unequal number of features, clustering results are strongly influenced by different initial values. Therefore, it is necessary to find the optimal sub-space set. To this end, a search algorithm is designed to find the optimal sub-space set. The search process is initialized by forming several sets (we call each set an environment) consisting of several numbers of sub-spaces. An optimal environment is the one that has the best clustering results. The search algorithm utilized three search operators to find the optimal environment. The search operators search all the environments and the consequent sub-spaces both locally and globally. These operators combine two environments and/or replace an environment with a newly generated one. Each search operator tries to find the best possible environment in the entire search space or in a local space. We evaluate the performance of our proposed clustering schema on 20 cancer gene datasets. The normalized mutual information (NMI) criterion and the adjusted rand index (ARI) are used to evaluate the performance evaluation. We first examine the effect of a different number of pairwise constraints. As expected, with increasing the number of pairwise constraints, the efficiency of the proposed method also increases. For example, the NMI value increases from 0.6 to 0.9 on the Khan-2001 dataset, when the number of pairwise constraints increases from 20 to 100. More number of pairwise constraints means more information is available, which helps to improve the performance of the clustering algorithm. Furthermore, we examine the effect of the number of random subspaces. It is observed that increasing the number of random subspaces has a positive effect on clustering performance with respect to the NMI value. In most datasets, when the number of sub-spaces reaches 20, the performance of the proposed method does not change much and is stable. Examining the effect of sampling rate for random subspace generation shows that the proposed method has the best performance in most cancer datasets, such as Armstrong-2002-v3, and Bredel-2005 datasets, when the random subspace generation rate is 0.5, and by deviating the rate from 0.5, the level of satisfaction decreases. Then, the results of the proposed idea are compared with the results of the method proposed in the reference [21] according to ARI and we see that our proposed method has performed better in 12 data sets out of 20 data sets than the method proposed in the reference [21]. Finally, the proposed idea is compared with some metric learning approaches with respect to NMI. We have observed that the proposed method obtained the best results compared to other compared methods on 11 datasets out of 20 datasets. It also achieved the second-best result on 6 out of 20 datasets. For example, the value NMI obtained in the proposed method is 0.1042 more than the reference [21] and it is 0.1846 more than RCA and it is 0.4 more than ITML and also it is 0.468 more than DCA on the Bredel-2005 dataset. Utilizing ensemble clustering methods besides the confidence factor improves the ability of the proposed algorithm to achieve better results. Also, utilizing the transitive operators as well as the selection of random subspaces of unequal sizes play an important role in achieving better performance for the proposed algorithm. Using the p-Laplacian spectral clustering method produces a better, more balanced, and normal volume of clusters compared to the standard spectral clustering. Another effective approach to the performance of the proposed method is to use search operators to find the best subspace, which leads to better results.

Recommender system can be defined as a software that suggests the most appropriate and closest item to the user's taste. They work as a counselor, behaving in such a way to guide people in the discovery of products of interest.Nowadays A great number of recommendation methods are used to implement a recommender system, a group of these algorithms are called collaborative filtering. These methods use the similarity between users or the similarity between items according to their user rating patterns for generating recommendations. Collaborative filtering algorithms can recommend the user, interesting items which are not similar to items she has rated before. These recommendations are generated according to the preferences of users with similar taste to the target user.  Different similarity functions and metrics have been used to create the model or compute the similarity in collaborative filtering methods. The best method which generates the most relevant items is not always the same and it may change according to the available data of users and items, because each approach has particularities and depends on the context to be applied. Thus, it becomes a hard task for system designers to manually select an appropriate method among the techniques.This article proposes an approach based on genetic algorithm for rank aggregation of memory based collaborative filtering methods and chooses the most relevant recommendations generated by different similarity techniques to create a Top-N recommender system. In order to implement these techniques, in addition to computing the similarity between users, inferred trust is also computed to increase the amount of available information about relations between user interests. The final method proposes a combination of collaborative filtering techniques for each data set, which in addition to considering time limits, has an acceptable precision for making recommendations.The proposed method has been compared against memory based collaborative filtering methods and similar methods. Experiments were performed using 1M MovieLens and 100k MovieLens and HetRec2011 data sets. The results show that the methodology proposed in this paper performs better and has a higher precision in generating recommendations for users than any of similar algorithms.

Hyperspectral images due to simultaneous acquisition of data in more than hundreds narrow and close spectral bands, have a very high correlation bandwidth. Hence, in order to store in less storage space, higher transmission speed and less bandwidth, they need compression. Various lossless and lossy methods for compression are exist, that can be in the spatial domain or in the spectrum domain. But, regard to the importance of spectral information of hyperspectral images in remote sensing, this compression should be done by this condition that the spectral information of this kind of images is well preserved. Compression methods can be based on either the predictive function or using of a codebook, to compress information. Data compression can also be done based on transformation coding, which these transformations can be cosine functions (DCTs), wavelet functions (DWTs), or principal component analysis (PCAs). Of course, PCA-based compression is one of the most effective ways to eliminate image correlations and reduce their volume. Another extension is the method of using curve fitting, which is applied exclusively to compress hyperspectral images due to its effect on the image spectrum. This method uses the spectral signature of the each pixel of image to reduce the feature by finding the closest approximation function to express the curve and storing its coefficients as a new feature for reconstruction compressed data. By replacing these coefficients in the equation of approximation, spectrum reflection curve for each pixel can be reconstructed. This method has very good results in comparison with previous methods such as PCA, but in compression using this method, the SRC curve has been approximated in some points with distortion. In this paper, we tried to eliminate these distortions, by finding points which have distortion and Breakdown the SCR. On the other hand, by using the Savitsky-Golay smoothing filter we can also reduce distortion and increase the PSNR. Another way to eliminate or reduce this distortion described in this article is as follow: At the first the spectral signature of each pixel of the intended data is smoothed by a Savitsky-Golay smoothing filter and then by using a particular method is divided into adjoining adjacent spaces and then a curve is plotted for each slice of data. By choosing the best degree and window length for smoothing and selecting the best degree of numerator and denominator of function, the coefficients of the selected rational function are considered as new features of the image. By using the proposed method, in addition to eliminating the distortion, the PSNR level is became much higher and the reconstructed image quality is very close to the original image.

Data clustering is one of the main tasks of data mining, which is responsible for exploring hidden patterns in unlabeled data. Due to the complexity of the problem and the weakness of the basic clustering methods, today most of the studies are directed towards clustering ensemble methods. Although for most datasets, there are individual clustering algorithms that provide acceptable results, but the ability of a single clustering algorithm is limited. In fact, the main purpose of clustering ensemble is to search for better and more stable results, using the combination of information and results obtained from several initial clustering. In this paper, a clustering ensemble-based method will be proposed, which, like most evidence accumulation methods, has two steps: 1- building a simultaneous participation matrix and 2- determining the final output from the proposed participation matrix. In the proposed method, some other information will be used in addition to the clustering of the samples to construct the simultaneous participation matrix. This information can be related to the degree of similarity of the samples, the size of the initial clusters, the degree of stability of the initial clusters, etc. In this paper, the clustering problem is defined as an explicit optimization problem by the mixed Gaussian model and is solved using the simulated annealing algorithm. Also, an evolutionary method based on simulated annealing will be presented to determine the final output from the proposed simultaneous participation matrix. The most important part of the evolutionary method is to determine the objective function that guarantees the final output will be of high quality. The experimental results show that the proposed method is better than other similar methods in terms of different clustering quality evaluation criteria.

In special purpose circuits, the amount of energy consumed and the speed of operation are the main challenges. There are wide researches and methods to improve the performance of these types of circuits. One of these methods is to use a Residue Number System (RNS). In the RNS, there are a number of modules (channels) as a set to represent the number and perform parallel arithmetic operations. The most famous set is the 3-modlui set {2n-1, 2n, 2n +1}. The form of modules to the power of 2 makes it easier to perform binary computational operations. To use this system, you need to perform conversion operations from binary to residue (forward conversion) and residue to binary (reverse conversion). The greater the number of modules (channels) in the set, the higher the degree of parallelism of computational operations. In contrast, more complex forward and reverse conversion circuits are required. The overhead of conversion computing can reduce the efficiency of using this system, unless the number of consecutive operations is large enough to cover the conversion overhead time. In this paper, based on 3-moduli set {2n-1, 2n, 2n +1} evaluation, it was determined that for how many consecutive addition or multiplication operations, the use of RNS operations leads to greater speed. In this paper, we evaluate the carry propagation adder as the most popular adder and parallel prefix adder as the high speed adder. Also, the parallel block multiplier circuit was used to evaluate the multiplication operations. First, modular adder/multiplier, binary, and forward and reverse conversion circuits were implemented and synthesized. We used Synopsys Design Compiler, K-2015.06 version and 45nm technology. The results show that if the carry propagation adder is used, in modules with a width of more than 8 bits (n≥8), if the number of consecutive operations is at least 4, it will speed up the calculations. Likewise, in the multiplication operation and parallel prefix addition, the number of sequences is reduced to two.

Rapid development in the Internet and communications have led to dramatic growth in computer networks, network size, and data exchange, and this can pose harmful threats to the network. Intrusion detection systems play an important role in the security of Internet networks, which protects the privacy, integrity, and availability of the network by inspecting network traffic. Intrusion detection models in the field of network security are predictive models that are used to predict malicious data in networks and one of the most widely used models in intrusion detection systems is based on machine learning. The imbalance between the accuracy of detection and false alarm rate is one of the most important challenges in this regard. In this paper, meta-heuristic algorithms are used to increase searchability and machine learning method is used to increase computational power and classification. Therefore, in this study, an efficient model based on the gray wolf algorithm and random forest algorithm to identify the best set of traffic features to identify and prevent cyberattacks is presented. The gray wolf algorithm is used to find the best feature subset and the random forest is used to evaluate each subset. This algorithm has also been improved to increase gray wolf performance. The accuracy obtained for correct classification in the proposed method in the NSL-KDD data set. as shown in the result, the detection accuracy of the traditional and improved gray wolf method is obtained 97.14% and 98.97%, respectively, which is outperformed other methods.

In the Corona crisis, we face a wide range of thoughts, feelings, attitudes, and behaviors on social media. This data contains valuable information for responding to the crisis by the people and administrators. The goal of this study is to identify the characteristics of messages that lead to different emotional polarities. This study aims to investigate the information posted by Twitter, Instagram, and Telegram users and news related to the COVID-19 pandemic in Iran. The data extracted from social networks are focused on the period of January 21, to April 29, 2020, which were shared in Iran and in Persian. It should be noted that the data set and their labels were published by the Cognitive Sciences and Technologies Council (CSTC) in Iran. In this work, the content of each post was pre-processed. Pre-processing was performed by removing stop words, normalizing the words, tokenizing, and stemming. The emotion labels were based on plutchik’s model and included joy, trust, fear, surprise, sadness, anticipation, anger, disgust, stress, and other emotions. In this study, clustering algorithms were used to analyze social media posts. We applied a two-stage clustering method. The proposed clustering algorithm was a combination of self-organized neural network and K-means algorithms. According to our proposed algorithm, the data were clustered through SOM at first, the results of which provided the initial cluster centers for the K-means algorithm. Implementations were built in Python version 3.7 and MATLAB R2015a. Hazm Tools was used for pre-processing data, and clustering was done in MATLAB. The Davies-Bouldin clustering evaluation was applied to find the optimal number of clusters. This measure was calculated for the number of clusters in the range of 2-50 in the two-stage clustering method. The results showed that the optimal number of clusters was ten. Analysis of the results showed that posts related to health and culture with negative polarity led to negative emotions such as fear, hatred, sadness, and anger. Messages about people's emotional and improper functioning have led to feelings of sadness, fear, and stress, and reduced hope in society. The results revealed a strong correlation between anger and disgust. Also, a positive correlation between fear, stress, and sadness was observed. In order to reduce the negative feelings and to create a sense of trust in the authorities, we suggest clarifying about the corona pandemic

Wireless Body Area Network (WBAN) is a pioneer trend in healthcare technology. Since any cyber-attack on a WBAN could jeopardize the patient's health, securing the WBAN plays a crucial role in healthcare applications. An intrusion detection system (IDS), as a second-line defense, is one of the security methods in computer networks. In this paper, a new IDS has been presented which is able to detect denial of service (DoS) attacks in a WBAN. In the proposed IDS, a genetic algorithm is used to select features of collected data, in a way that increases the performance of the IDS and as a result the WBAN. Then, using support vector machine and k nearest neighbor techniques, the data classification is performed to detect DoS traffic from regular data traffic. Simulation results indicate that the proposed IDS has effective performance with a 90% detection rate.

In this paper, a new recurrent type-2 fuzzy neural network for nonlinear dynamic systems identification is presented. The structure of the new type-2 fuzzy neural network with the non-linear "then" part has 8 layers. In layers 0, 1 and 2, the fuzzification operation is performed and the upper and lower limits of the membership degree are determined. Normalization and weighting operations are performed in layers 3 and 4. In layer 5, there are non-linear trigonometric functions, which actually form the "then" part of the fuzzy system, and return feedback from the output layer enters this layer. Finally, in the 6th and 7th layers, the de-fuzzification operation and the output calculation are performed. In order to check and evaluate the performance of the network in system identification, the input-output information of two physical systems (a DC motor and a flexible robot arm) has been applied to the type-2 recurrent fuzzy neural network. This research is completely experimental and practical, in other words, it is the use of artificial intelligence techniques in operational work. Among the innovations of this article, in addition to presenting a new neural network, is generating a suitable signal to stimulate the system, extracting data from practical systems, data pre-processing (removing outliers, estimating missing data, and normalizing data). In the simulation, the root mean square error criterion shows that the proposed method has a better performance than other methods.

Finding a highly informative, low-dimensional representation for texts, specifically long texts, is one of the main challenges for natural language processing (NLP) tasks. For texts longer than sentences or a paragraph, finding a good representation beyond the bag-of-words model without losing word order is still a challenge. This representation should capture the semantic and syntactic information of the text while retaining relevance for large-scale similarity search and accurate text classification. We propose the utilization of Rhetorical Structure Theory (RST) to consider the text structure in the representation. RST creates a tree-structure format for the text document and model the importance and relationship between sentences or phrases. In this paper, we examine the effect of using this structure on two different NLP tasks. In information retrieval, to embed document relevance in distributed representation, we use a Siamese neural network to jointly learn document representations. Our Siamese network consists of two sub-networks of recursive neural networks (RNN) built over the RST tree. For this task, we use a subset of Reuters’s news corpus and BBC news dataset. The results show that our approach outperforms conventional text representations like tf_idf, LDA, LSA and word vector averaging. The proposed representation beats the best conventional method by %6 and %3 in precision at k retrieved documents on BBC and Reuters datasets, respectively. In the sentiment analysis task, first, we use an rst-based recursive neural network to represent movie reviews and classify the polarity of people’s opinions. Then we propose to use the nucleus-satellite information of a node in the rst-tree to build an attention mechanism by deep RNN to generate better discourse representations. We test the effectiveness of our approach on  sentiment analysis task, and we prove that considering the importance of the text span improves sentiment analysis performance by %3 on the internet movie review database. In this paper, we improve the text representation by the rst-based deep neural network. We can evaluate this approach on the other languages to show the effectiveness of using the structure format of the text.