فصلنامه پردازش علائم و داده ها
سال نوزدهم شماره 2 (پیاپی 52، تابستان 1401)

Potholes on roads are regarded as serious problems in the transportation domain, and ignoring them lead to an increase in accidents, traffic, vehicle fuel consumption, and waste of time and energy. As a result, pothole detection has attracted researchers’ attention, and different methods have been presented for it up to now. Data analysis methods such as machine learning and soft computing have been widely used for detection purposes. They rely on a dataset and propose a system that can detect a special event in similar datasets. Their effectiveness can be measured by evaluating their accuracy in detecting the event. Image processing involves a wide range of analytics that are used to extract specific information from images. The majority of image processing programs require massive computational power. The major part of previous research is based on image processing. They utilize dedicated cameras which are embedded in vehicles to take images and analyze them through massive image processing programs. This scheme requires dedicated hardware that is not typically available on vehicles. In this paper, a new scheme is proposed, which uses accelerometer and GPS sensors. These types of sensors are available in today’s smartphones as well as modern vehicles. The data generated by these sensors is processed via soft computing to increase the accuracy of pothole detection. The proposed algorithm uses a combination of a fuzzy system and evolutionary algorithms. Fuzzy systems have been widely used to model the real-world problems that are described by uncertainty and ambiguity. Evolutionary algorithms (e.g., genetic algorithms) try to imitate evolutionary science in solving hard problems. Genetic algorithm and harmony search are used to adjust membership functions of the proposed fuzzy system. For evaluation, a case study has been conducted with regard to detect potholes on Ghaffari Street in Birjand. To this end, a real dataset has been collected and used for implementing the proposed method. Experimental results show the high accuracy of the proposed algorithm in comparison to other solutions. They reveal that the accuracy of the proposed genetic fuzzy algorithm is 98 percent and for the proposed harmony fuzzy algorithm is 99 percent.

Orthogonal Frequency Division Multiplexing (OFDM) is a useful technology in wireless communications that provides high-rate data transmission in multipath fading channels. The advantages of OFDM systems are the high spectral efficiency and strong resistance to frequency selective fading. In OFDM systems, a large number of sub-carriers are used to modulate the symbols causing the time-domain OFDM signal to have a large dynamic range, or a high peak-to-average power ratio (PAPR). When the signals are applied to a nonlinear power amplifier, the OFDM systems’ performance is degraded by the high PAPR. In recent years, several works have been done to reduce the PAPR of OFDM systems. One of the most well-known methods is a partial transmit sequence (PTS). Regardless of the PTS advantages, it suffers from a high computational complexity. Because it requires an exhaustive search over all possible combinations of phase factors. The computational complexity of the PTS increases with increasing the number of phase factors and sub-blocks. There are several approaches to overcome the computation complexity issue of the PTS technique. The majority of these methods mainly employed swarm intelligence and evolutionary optimization algorithms to resolve the PTS shortcoming. These methods report encouraging results, however, their performance is far from the ideal state. This highlights that improving the performance of PTS is an open problem and there is room for more improvement. As an element of research, we propose an optimization approach based on the election algorithm (EA) to overcome the computational complexity of the PTS technique. To realize this goal, we improve the EA algorithm by introducing a new version of positive advertisements operator. The new operator efficiently improves the search capability of the EA through balancing between the exploration and exploitation power of the algorithm. The proposed EA based PTS (EA-PTS) approach, by searching the optimal phase factors, imposes less computational complexity on the system and reduces the PAPR to an acceptable level. The proposed method is compared with the optimal PTS (O-PTS), genetic algorithm-based PTS (GA-PTS) and imperialism competition algorithm based PTS (ICA-PTS) techniques. Simulation results show that the proposed EA-PTS has better performance in simultaneously reducing the PAPR and computational complexity.

One of the biggest challenges in this field is classification problems which refers to the number of different samples in each class. If a data set includes two classes, imbalance distribution occurs when one class has a large number of samples while the other is represented by a small number of samples. In general, the methods of solving these problems are divided into two categories: under-sampling and over-sampling. In this research, it is focused on under-sampling and the advantages of this method will be analyzed by considering the efficiency of classifying imbalanced data and it’s supposed to provide a method for sampling a majority data class by using subtractive clustering and fuzzy similarity measure. For this purpose, at first the subtractive clustering is conducted and the majority data class is clustered. Then, using fuzzy similarity measure, samples of each cluster will be ranked and appropriate samples are selected based on these rankings. The selected samples with the minority class create the final dataset. In this research, MATLAB software is used for implementation, the results are evaluated by using AUC criterion and analyzing the results has been performed by standard statistical tools. The experimental results show that the proposed method is superior to other methods of under-sampling.

Multivariate time series data are found in a variety of fields such as bioinformatics, biology, genetics, astronomy, geography and finance. Many time series datasets contain missing data. Multivariate time series missing data imputation is a challenging topic and needs to be carefully considered before learning or predicting time series. Frequent researches have been done on the use of different techniques for time series missing data imputation, which usually include simple analytic methods and modeling in specific applications or univariate time series. In this paper, a hybrid approach to obtain missing data is proposed. An improved version of inverse distance weighting (IDW) interpolation is used to missing data imputation. The IDW interpolation method has two major limitations: 1) finding closest points to missing data 2) Choosing the optimal effect power for missing data neighbors. Clustering has been used to remove the first constraint and find closest points to the missing data. With the help of clustering, the search radius and the number of input points that are supposed to be used in interpolation calculations are limited and controlled, and it is possible to determine which points are used to determine the value of a missing data.Therefore, most similar data to the missing data are found. In this paper, the k-maens clustering method is used to find similar data. This method has been more accurate than other clustering methods in multivariate time series. Evolutionary algorithms are used to find the optimal effect power of each data point to remove the second constraint. Considering that each sample within each cluster has a different effect on the estimation of missing data, cuckoo search is used to find the effect on missing data. The cuckoo search algorithm is applied to the data of each cluster, and each data sample that has more similarity with the missing data has more influence, and each data sample that has less similarity has less influence and has less influence in determining the amount of missing data. Among evolutionary algorithms, evolutionary cuckoo search algorithm is used due to high convergence speed, much less probability of being trapped in local optimal points, and ability to quickly solve high dimensional optimization problems in multivariate time series problems. To evaluate the performance of the proposed method, RMS, MAE, , MSE and MAPE criteria are used. Experimental results are investigated on four UCI datasets with different percentages of missingness and in general, the proposed algorithm performs better than the other three comparative methods with an average RMSE error of 0.05, MAE error of 0.04, MSE error of 0.003, and MAPE error of 5. The correlation between the actual data and the estimated value in the proposed method is about 99%.

In recent years, Vehicular Ad-hoc Networks (VANET) as an emerging technology have tried to reduce road damage and car accidents through intelligent traffic controlling. In these networks, the rapid movement of vehicles, topology dynamics, and the limitations of network resources engender critical challenges in the routing process. Therefore, providing a stable and reliable routing algorithm is a necessary requirement to maintain the Quality of Service (QoS) parameters of VANETs. In this paper, a new routing algorithm based on the clustering technique is proposed, which is called GCAR. In the proposed algorithm, the appropriate cluster heads are selected based on the genetic algorithm then two vehicles are selected between the neighboring clusters as the gateways and a vehicle chain is formed by these vehicles. Moreover, a combination of genetic algorithm and simulated annealing is applied to identify the suitable clusters. The conducted simulations in MATLAB tool indicate that, respectively, path discovery ratio, the number of clusters, throughput, and packet delivery ratio of the proposed algorithm have been improved by 18.4%, 2.55%, 3.45%, and 14.18% in comparison to the PassCAR approach. Furthermore, evaluation of the convergence, standard deviation, and standard error of the proposed algorithm prove its high convergence speed and stability.

Today, we are witnessing the expansion of various Internet of Things (IoT) applications and services such as monitoring and health. These services are delivered to users via smart devices anywhere and anytime. Forecasts show that the IoT, which is controlled online in the user environment, will reach 25 billion devices worldwide by 2020. Data security is one of the main concerns in the IoT. The IoT is supposed to deal with a population of about billions of objects, so the number of malicious attacks can be very high and alarming given the global connection (anyone access) and the wide availability (access to any place at any time). However, these accesses can make security and privacy critical. Reports show that 26% of IoT attacks in 2019 were related to non-authentication, which is why IoT authentication has become one of the most sensitive security concepts. IoT devices are usually left unattended and this makes it easy for an attacker to target such equipment. For example, security breaches and unwanted changes in patient's health parameters in smart health care systems can cause wrong treatments or even lead to his death. The fact that each device in the IoT knows who it is communicating with and at what level of access is one of the important aspects of security, especially in cases where various devices with different capabilities have to perform common tasks and cooperate with each other. IoT authentication is a trust model to protect control access and data when information travels between devices. So far, different methods have been proposed for authentication in the IoT network. These methods are usually based on the public key, private key, random key distribution, and hash function. A point that should be taken into account in IoT authentication is that IoT networks and devices have limited bandwidth, low memory, low processing power, and energy limitations. Therefore, the proposed method should pay special attention to such limitations. In addition, IoT authentication needs to ensure enhanced security features such as confidentiality, data integrity, reliability, maintainability, scalability, and privacy to their consumers. This paper proposes a two-way or mutual authentication protocol in which both devices authenticate each other without human intervention in a smart home network. The proposed protocol is based on asymmetric encryption for authentication of devices, which have a shared private session key, along with hashing operations in the network. Also, to ensure the security of communications at each session, each device has a one-time private session key. The session keys are changed regularly to ensure the security of sessions between devices. The proposed protocol is programmed by HLPSL and simulated and verified by the SPAN and AVISPA tools. The security analysis results show the proposed protocol is extremely practical, and secure against potential attacks.

Optimization is a very important process in engineering. Engineers can create better production only if they make use of optimization tools in reduction of its costs including consumption time. Many of the engineering real-word problems are of course non-solvable mathematically (by mathematical programming solvers). Therefore, meta-heuristic optimization algorithms are needed to solve these problems. Based on this assumption, many new meta-heuristic optimization algorithms have been proposed inspired by natural phenomena, such as IWO [58], BBO [59], WWO [61], and so on. Inspired by domino toppling theory, we proposed an optimization algorithm. Using domino pieces, we can create countless complex structures. To simulate the domino movement in the search space of a problem, we consider the particles in the search space as the domino pieces and, by creating an optimal path, we will try to direct the dominoes to the optimal path. The optimal paths will be updated in each iteration. After initializing the dominoes randomly at the beginning of each evaluation, the picking piece or the first moving piece will be identified and then the particles will be selected by the optimal path. Applying a motion equation to each domino will move the dominoes forward in that direction. At first, a predefined dominoes will be randomly distributed in the problem space. Choosing the optimal path will accelerate the convergence of the domino particles towards the target. After choosing the path in current iteration, we now have to do the domino movement. The particles will move to a new location by applying the new location equation. By applying this equation, each domino piece will sit on the track ahead of itself. The front piece will also move to a new location by applying an equation separate from the rest. After moving the dominoes to the new location, the worst iteration of the previous iteration will be removed from the problem space. In the new iteration, the optimal domino path, the new locations of domino pieces and the global optimum will be updated. At the end of the algorithm, the global optimum will be determined as the optimal solution. This method is implemented in a simulator environment. To evaluate the performance of the Domino Optimization algorithm, we use a complete benchmark including 30 objective functions called CEC 2014 [67] that are single-objective numerical functions. In all cases, we set the population size to 50, the dimension size to 30, and the number of fitness function evaluation to 150,000. We compare the proposed Domino Optimization algorithm (DO) with the algorithms LOA [57], ICS [62], NPSO [63], MOHS [64], BCSO [65] and FFFA [66]. The results obtained from the 3 unimodal functions show that the proposed method is able to achieve a better solution than any of the state of the art algorithms at the equal resources. Results in the multimodal functions show that the proposed method has the best performance in finding the optimal solution in all of the available 13 functions in this section. In all of 6 functions in the hybrid section, the quality of the proposed method is better than all of the state of the art algorithms at the equal resources. The standard deviation values ​​of the proposed method, which are often small numbers, indicate algorithm convergence around the optimal solution. Also among the available methods, two algorithms, named NPSO and LOA, have good results after the proposed method. In the convergence analysis of dominoes, the diversity of objective functions in 100 distinct iterations shows a big value at the beginning of the algorithm, and a low value at the end of the algorithm.

With the explosive growth of social media such as Twitter and Instagram, reviews on e-commerce websites, and comments on news websites, individuals and organizations are increasingly using analyzing opinions in these media for their decision-making and designing strategies. Sentiment analysis is one of the techniques used to analyze users' opinions in recent years. The Persian language has specific features and thereby requires unique methods and models to be adopted for sentiment analysis, which are different from those in English and other languages. This paper identifies the characteristics and limitations of the Persian language. Sentiment analysis in each language has specified prerequisites; hence, the direct use of methods, tools, and resources developed for the English language in Persian has its limitations. The present study aims to investigate and compare previous sentiment analysis studies on Persian texts and describe views presented in articles published in the last decade. First, the sentiment analysis levels, approaches, and tasks are described. Then, a detailed survey of the applied sentiment analysis methods used for Persian texts is presented, and previous works in this field are discussed. The advantages and disadvantages of each proposed method are demonstrated. Moreover, the publicly available sentiment analysis resources of Persian texts are studied, and the characteristics and differences of each are highlighted. As a result, according to the recent development of the sentiment analysis field, some issues and challenges not being addressed in Persian texts are listed, and some guidelines are provided for future research on Persian texts. Future requirements of Persian text for improving the sentiment analysis system are detailed.

Extracting structured information about entities from web texts is an important task in web mining, natural language processing, and information extraction. Information extraction is useful in many applications including search engines, question-answering systems, recommender systems, machine translation, and etc. An information extraction system aims to identify the entities from the text and extract their related information to form a profile of the target entity. In recent years, several methods have been proposed for extracting structured information from web text. The majority of existing methods for extracting entity-centric information require a predefined ontology. The ontology includes the complete knowledge of the entities and their attributes. The main challenge of these methods is their inability to extract information about entities that are not already defined in the ontology. Besides, the existing methods have ignored semantic information extraction and have not linked the extracted information to the general ontology entries. This highlights that introducing new methods for semantic information extraction is an open problem and there is room for more efforts in this field. As an element of research, we proposed a new method for the automatic extraction of semantically structured information from Farsi web text. The proposed method does not require background knowledge about the entities and their properties. The proposed method consists of three main phases including pre-processing, semantic analysis and frame extraction. To fulfill these phases, we use a combination of language resources, text processing tools, and distant ontologies. The main focuses of the proposed method are to enrich the predicate-argument frames with the semantic information extracted from distant ontologies, extract the entity-related information from predicate-argument frames, and link the extracted information with their corresponding sense in DBPedia ontology. The issue facilitates the processing of Farsi texts by computers. To evaluate the proposed method, we created a small Farsi dataset containing 100 complete sentences. Then, the proposed method is compared with three information extraction methods on this dataset. The results of experiments show the superiority of the proposed method compared to counterpart methods in terms of precision and F1 measures.

In the past years, a number of algorithms have been introduced for synthesis aperture radar (SAR) imaging. However, they all suffer from the same problem: The data size to process is considerably large. In recent years, compressive sensing and sparse representation of the signal in SAR have gained a significant research interest. This method offers the advantage of reducing the sampling rate but also suffers from speed processing limitation and it needs a huge amount of memory to reconstruct the image. On the other hand, inaccuracy in SAR model induces phase error to the results and makes the reconstructed image blurry. Existing sparse methods in the presence of phase error, have high computational costs and need a lot of processing time. In addition, these methods take up considerable space in the memory for saving the measurement matrix. In this paper, a fast method is proposed to reduce the computational cost of image reconstruction, based on the signal sparsity in the presence of phase error. The proposed method consists of substituting accurate observations of sparsity methods with approximated observations of matched filter methods. In this method, the output of Range-Doppler matched filter is reconstructed with sparse representation, and error phase is estimated simultaneously. This method leads to a nonconvex optimization problem and to solve that, we use the majorization minimization method. The phase error and reconstructed image are estimated in an iterative procedure. The use of approximated observation, eliminates the need for carrying out big matrix multiplications, and Fast Fourier Transformation, as a low computational cost operation, can be employed instead. In addition to computation speed, this method does not need any memory space for saving measurement matrices. In our numerical simulations, we compared the speed of processing and the mean square error (MSE) of reconstructed images for the proposed method with the state-of-the-art sparse method for different sizes of image and under-sampling rates. It is shown in simulations that the reconstructed image from our method has a slightly lower quality and higher MSE, because of the sidelobes effect of the matched filter output. However, in certain conditions, the speed of the proposed method is more than a hundred times faster than the compared method. The achieved processing speed with no need for the memory to store the measurement matrix at the expense of slightly lower image quality would be acceptable for most applications.

One of the most important research areas in natural language processing is Question Answering Systems (QASs). Existing search engines, with Google at the top, have many remarkable capabilities. However, there is a basic limitation; search engines do not have deduction capability which a QAS is expected to have. In this perspective, a search engine may be viewed as a semi-mechanized QAS. Upgrading a search engine such to a QAS is a task whose complexity is hard to exaggerate. To achieve success, new concepts and ideas are needed to address difficult problems which arise when knowledge has to be dealt with in an environment of imprecision, uncertainty and partial truth.  QASs are search engines that have the ability to provide a brief and accurate answer to each question in natural language for instance, the question that a search engine answers with a set of documents, a QAS answers with a paragraph, sentence or etc. In this paper, a solution is proposed to optimize the performance and speed of web-based QASs for answering English questions.  As evolutionary algorithms are suitable for issues with large search space, in this approach we have used an evolutionary algorithm to optimize QASs. In this regard, we have chosen APSO which is a simplified version of PSO. The proposed method consists of five main stages: question analysis, pre-process, retrieval, extraction and ranking. We have tried to provide a method that would be more accurate in choosing the most probable answer from the documents that have been retrieved by the standard search engine and at the same time, be faster than similar methods. In ranking process, various attributes can be extracted from the text that are used in APSO. For this purpose, in addition to selecting a sentence from the text and examining its attributes, different cut parts of the sentence are selected each time by changing the beginning and end points of the cut part. The attributes which have been used in this study are: 1. Number of unigrams similar to the question words, 2. Number of bigrams similar to the question words, 3. Number of unigrams similar to the question words in the cut part, 4. Number of bigrams similar to the question words in the cut part, 5. Number of synonyms with the question words and 6. Number of synonyms with the question words in the cut part. The fitness function is the weighted sum of these attributes. Top-1 accuracy and MRR are the most valid metrics for measuring the performance of QASs. The proposed method has achieved the accuracy (top-1 accuracy) of 0.527 with respect to the standard dataset and the MRR of it, is 0.711. Both of these results are improved compared to most similar systems. In addition, the time taken to answer the input question in the proposed method, has been significantly reduced compared to similar methods. In general, the accuracy and MRR in this paper have progressed and the system needs less time to find the answer, in comparison with existing QASs.

Nowadays, the use of various messaging services is expanding worldwide with the rapid development of Internet technologies. Telegram is a cloud-based open-source text messaging service. According to the US Securities and Exchange Commission and based on the statistics given for October 2019 to present, 300 million people worldwide used telegram per month. Telegram users are more concentrated in countries such as Iran, Venezuela, Nigeria, Kenya, Russia, and Ukraine. This messenger has become a popular and extensively used messenger because it supports various languages and provides diverse services such as creating groups and channels with a large number of users and members. There is a large amount of contextual data on telegram groups containing hidden knowledge; the extraction of this knowledge can be beneficial. The requests on telegram users' messages are examples of this sort of data with hidden knowledge. Hence, identifying requests can respond to users' needs and help them fulfill their desires immediately; this drives users' business development. The authors identified these requests in a telegram search engine named the Idekav system of Yazd University. Then, the authors created opportunities to earn money by sending these requests to the business owners who were able to respond to them. Given the high dimensions of feature space in contextual data, it is necessary to reduce attributes using feature selection.        In the present study, the appropriate features were selected for Persian text classification and request identification. Among the feature selection methods, two local and global filter-based methods were chosen. By general investigation and combining the most extensively used filter-based FS methods, an optimal subset of important features was obtained. This hybrid feature selection method resulted in increased request identification accuracy, improved Persian text classification efficiency, and reduced training time and computation by optimizing the feature reduction. Of course, it is noteworthy that the classification accuracy is reduced in some methods; however, this value is negligible compared to the feature reduction value. Incorporating the concept of opinion mining into the analysis of emotions and questions can be a method to identify positive or negative demand in social networks. Therefore, the requests in the Persian telegram messages can be identified using opinion mining researches. For experiments in the present article, a dataset called Persian is used, which is extracted from the Idekav system. The selection of suitable features to increase model accuracy in request identification is an important part of this research. The support vector machine was employed to calculate accuracy. Given the acceptable results of the SVM, its various kernels were also calculated. Micro-averaging and macro-averaging criteria were also used for evaluation. Model inputs include many optimal feature subsets. Furthermore, feature selection methods have been proposed to produce suitable features for each model for increasing the accuracy of the model. Afterward, among all the features investigated, appropriate features have been selected for each of the applied feature selection models. For a more precise explanation, the main innovations of the present study are as follows: Use of the most common filters based on local and global feature selection methods to find the optimal feature set. Use of hybrid methods to create suitable features for predictive models of accuracy in Persian text classification and their application in identifying requests in Persian messages on telegram. Selecting suitable features to increase accuracy and reduce computational time for each of the models under consideration. In this regard, in addition to picking an efficient algorithm, it is attempted to provide a method for making more appropriate choices. Evaluation and testing of the proposed models for a large set of Persian data and many different features.