Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420542
Shakiba Tasharrofi, H. Taheri
Neural networks had impressive results in recent years. Although neural networks only performed using Euclidean data in past decades, many data-sets in the real world have graph structures. This gap led researchers to implement deep learning on graphs. The graph convolutional network (GCN) is one of the graph neural networks. We propose the differential evolutional optimization method as an optimizer for GCN instead of gradient-based methods in this work. Hence the differential evolution algorithm applies for graph convolutional network’s training and parameter optimization. The node classification task is a non-convex problem. Therefore DE algorithm is suitable for these kinds of complex problems. Implementing evolutionally algorithms on GCN and parameter optimization are explained and compared with traditional GCN. DE-GCN outperforms and improves the results by powerful local and global searches. It also decreases the training time.
{"title":"DE-GCN: Differential Evolution as an optimization algorithm for Graph Convolutional Networks","authors":"Shakiba Tasharrofi, H. Taheri","doi":"10.1109/CSICC52343.2021.9420542","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420542","url":null,"abstract":"Neural networks had impressive results in recent years. Although neural networks only performed using Euclidean data in past decades, many data-sets in the real world have graph structures. This gap led researchers to implement deep learning on graphs. The graph convolutional network (GCN) is one of the graph neural networks. We propose the differential evolutional optimization method as an optimizer for GCN instead of gradient-based methods in this work. Hence the differential evolution algorithm applies for graph convolutional network’s training and parameter optimization. The node classification task is a non-convex problem. Therefore DE algorithm is suitable for these kinds of complex problems. Implementing evolutionally algorithms on GCN and parameter optimization are explained and compared with traditional GCN. DE-GCN outperforms and improves the results by powerful local and global searches. It also decreases the training time.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129584891","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420556
Amir Soltany Mahboob, M. R. Moghaddam
Neuro-fuzzy systems have been proved effective in training classifiers, especially when it comes to noisy, inaccurate or incomplete datasets. For this reason, and due to their simple comprehensible nature, these systems have become popular in designing classifiers. One of the major challenges in designing a neuro-fuzzy classifier is achieving the optimum system parameters such as the type and position of the membership function as well as its training method. These factors could affect the function of the classifier significantly. In this paper, a novel method based on evolutionary algorithms such as inclined planes optimization algorithm (IPO), particle swarm optimizer (PSO) and genetic algorithm (GA) is introduced to design a neuro-fuzzy classifier in such a way that the accuracy is increased and the error rate is minimized. To prove the efficiency of the proposed method, several experiments are conducted on well-known datasets with different number of classes and different feature vector lengths. Results indicate that the proposed evolutionary-based neuro-fuzzy classifier is superior to a normal neuro-fuzzy classifier in terms of accuracy. In addition, experiments showed that the proposed method is able to properly classify the data with a relatively high stability.
{"title":"A Neuro-Fuzzy Classifier Based on Evolutionary Algorithms","authors":"Amir Soltany Mahboob, M. R. Moghaddam","doi":"10.1109/CSICC52343.2021.9420556","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420556","url":null,"abstract":"Neuro-fuzzy systems have been proved effective in training classifiers, especially when it comes to noisy, inaccurate or incomplete datasets. For this reason, and due to their simple comprehensible nature, these systems have become popular in designing classifiers. One of the major challenges in designing a neuro-fuzzy classifier is achieving the optimum system parameters such as the type and position of the membership function as well as its training method. These factors could affect the function of the classifier significantly. In this paper, a novel method based on evolutionary algorithms such as inclined planes optimization algorithm (IPO), particle swarm optimizer (PSO) and genetic algorithm (GA) is introduced to design a neuro-fuzzy classifier in such a way that the accuracy is increased and the error rate is minimized. To prove the efficiency of the proposed method, several experiments are conducted on well-known datasets with different number of classes and different feature vector lengths. Results indicate that the proposed evolutionary-based neuro-fuzzy classifier is superior to a normal neuro-fuzzy classifier in terms of accuracy. In addition, experiments showed that the proposed method is able to properly classify the data with a relatively high stability.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130019226","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420592
Amene Naghdipour, S. Hasheminejad
The software design phase is important and challenging due to its high impact on other phases of the software development life cycle. Design patterns are proven solutions based on software developers' experience to solve recurring problems, which used to acquire quality software design. However, selecting an appropriate design pattern is quite difficult. Hence, many studies have been done to automate the design pattern selection process. The existing automated design pattern selection methodologies have certain issues such as the need to have a large sample size, user restrictions on selecting preset concepts, time-consuming, and incomprehensiveness. To address these issues in this paper, a two-phase method for selecting an appropriate design pattern is presented. The proposed method is based on an ontology approach that enables domain knowledge to be modeled in a simple and abstract way and enables queries to be evaluated against a knowledge base. The concepts of ontology are then linked to WordNet. Subsequently, a dataset includes use cases that can be satisfied with GOF design patterns is provided. The set of use cases is then processed in such a way as to make it easy and fast to select the concept-constraint pair to query the ontology. The experimental shows promising and effective results of the proposed method.
{"title":"Ontology-Based Design Pattern Selection","authors":"Amene Naghdipour, S. Hasheminejad","doi":"10.1109/CSICC52343.2021.9420592","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420592","url":null,"abstract":"The software design phase is important and challenging due to its high impact on other phases of the software development life cycle. Design patterns are proven solutions based on software developers' experience to solve recurring problems, which used to acquire quality software design. However, selecting an appropriate design pattern is quite difficult. Hence, many studies have been done to automate the design pattern selection process. The existing automated design pattern selection methodologies have certain issues such as the need to have a large sample size, user restrictions on selecting preset concepts, time-consuming, and incomprehensiveness. To address these issues in this paper, a two-phase method for selecting an appropriate design pattern is presented. The proposed method is based on an ontology approach that enables domain knowledge to be modeled in a simple and abstract way and enables queries to be evaluated against a knowledge base. The concepts of ontology are then linked to WordNet. Subsequently, a dataset includes use cases that can be satisfied with GOF design patterns is provided. The set of use cases is then processed in such a way as to make it easy and fast to select the concept-constraint pair to query the ontology. The experimental shows promising and effective results of the proposed method.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133519458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420631
S. Araei, A. Ghomsheh
3D action recognition is a long-standing problem in the field of computer vision. Given the 3D coordinate set of body joints, it is desired to recognize what activity is performed. The problem can be approached using a time-series model. Recent advancements in the field of recurrent neural networks have enabled the use of sophisticated memory cells that can predict time series using the information from earlier elements of a sequence. In this article, we proposed a hierarchical architecture that attends to its own signature through time, which can put more weight on time frames of the sequence that are more specific to the performed action. Accordingly, using memory cells, a self-attention mechanism is implemented. In addition, spatial attention is also considered by sub-grouping and then regrouping body parts down the architecture hierarchy. We evaluate the proposed model on NTU and MSR 3D action datasets. An accuracy of 79.8% and 97.8% on NTU and MSR datasets indicated that the proposed method outperforms the previous methods tested in this paper.
{"title":"Spatio-Temporal 3D Action Recognition with Hierarchical Self-Attention Mechanism","authors":"S. Araei, A. Ghomsheh","doi":"10.1109/CSICC52343.2021.9420631","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420631","url":null,"abstract":"3D action recognition is a long-standing problem in the field of computer vision. Given the 3D coordinate set of body joints, it is desired to recognize what activity is performed. The problem can be approached using a time-series model. Recent advancements in the field of recurrent neural networks have enabled the use of sophisticated memory cells that can predict time series using the information from earlier elements of a sequence. In this article, we proposed a hierarchical architecture that attends to its own signature through time, which can put more weight on time frames of the sequence that are more specific to the performed action. Accordingly, using memory cells, a self-attention mechanism is implemented. In addition, spatial attention is also considered by sub-grouping and then regrouping body parts down the architecture hierarchy. We evaluate the proposed model on NTU and MSR 3D action datasets. An accuracy of 79.8% and 97.8% on NTU and MSR datasets indicated that the proposed method outperforms the previous methods tested in this paper.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"129 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122451326","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420555
Zahra Nekudari, A. Ghasemi
Algebraic connectivity is a global criterion for assessing network resistance to failures. Algebraic connectivity is a monotone measure against the number of links added to a given network to enhance its robustness. In this paper, we show the effect of link addition on the size of cascading failures. Accordingly, we consider two different strategies for step-by-step link addition: adding links to the network’s core and adding links to the whole network. We choose new links using simulated annealing to maximize the algebraic connectivity. Simulation results suggest that although the core of the network has a significant impact on network robustness, adding links to the core did not significantly affect cascading failures. Conversely, we find that adding links to the whole network make the network robust against cascading failures.
{"title":"The effect of increasing the algebraic connectivity on cascading failures in power grid networks","authors":"Zahra Nekudari, A. Ghasemi","doi":"10.1109/CSICC52343.2021.9420555","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420555","url":null,"abstract":"Algebraic connectivity is a global criterion for assessing network resistance to failures. Algebraic connectivity is a monotone measure against the number of links added to a given network to enhance its robustness. In this paper, we show the effect of link addition on the size of cascading failures. Accordingly, we consider two different strategies for step-by-step link addition: adding links to the network’s core and adding links to the whole network. We choose new links using simulated annealing to maximize the algebraic connectivity. Simulation results suggest that although the core of the network has a significant impact on network robustness, adding links to the core did not significantly affect cascading failures. Conversely, we find that adding links to the whole network make the network robust against cascading failures.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"92 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115879113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420561
Mohammadhossein Kiyaei, Farkhondeh Kiaee
ATMs are no longer just machines, these connected devices are smart, intelligent things in the Internet of Things (IoT). Access to cash for many in society is remaining essential during the current COVID-19 lock-down around the globe. A cash inventory management system is necessary to decide whether ATM should be replenished on each day of the week. In this paper, we study the real-time cash replenishment planning problem under outflow uncertainty where the fee of the security companies grows if the replenishment ends up falling on a weekends/holidays. Our model is based by the Double Deep Q-Network (DQN) algorithm which combines popular Q-learning with a deep neural network. The proposed method is used to control replenishment operation in order to minimize replenishment cost where the cash demand changes dynamically at each day. Experiment results show that our proposed method can work effectively on the real outflow time-series and it is able to reduce the ATM operational cost compared with the other state-of-the-art cash demand prediction schemes.
{"title":"Optimal ATM Cash Replenishment Planning in a Smart City using Deep Q-Network","authors":"Mohammadhossein Kiyaei, Farkhondeh Kiaee","doi":"10.1109/CSICC52343.2021.9420561","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420561","url":null,"abstract":"ATMs are no longer just machines, these connected devices are smart, intelligent things in the Internet of Things (IoT). Access to cash for many in society is remaining essential during the current COVID-19 lock-down around the globe. A cash inventory management system is necessary to decide whether ATM should be replenished on each day of the week. In this paper, we study the real-time cash replenishment planning problem under outflow uncertainty where the fee of the security companies grows if the replenishment ends up falling on a weekends/holidays. Our model is based by the Double Deep Q-Network (DQN) algorithm which combines popular Q-learning with a deep neural network. The proposed method is used to control replenishment operation in order to minimize replenishment cost where the cash demand changes dynamically at each day. Experiment results show that our proposed method can work effectively on the real outflow time-series and it is able to reduce the ATM operational cost compared with the other state-of-the-art cash demand prediction schemes.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116403324","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420597
Ramin Fotouhi, M. Pourgholi
This paper presents an approach for controlling the multi-agent system based on optimal control approach. The cost function of this problem is global, and three algorithms (Jaya, teaching-learning and water cycle algorithms) are applied to the system. Simulation outputs show the usefulness of the water cycle algorithm so as to find the better performance in terms of complexity of algorithm for the problem, and this technique leads to optimal consensus. Simulations are done via Matlab software.
{"title":"Water Cycle Algorithm-Based Control for Optimal Consensus Problem","authors":"Ramin Fotouhi, M. Pourgholi","doi":"10.1109/CSICC52343.2021.9420597","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420597","url":null,"abstract":"This paper presents an approach for controlling the multi-agent system based on optimal control approach. The cost function of this problem is global, and three algorithms (Jaya, teaching-learning and water cycle algorithms) are applied to the system. Simulation outputs show the usefulness of the water cycle algorithm so as to find the better performance in terms of complexity of algorithm for the problem, and this technique leads to optimal consensus. Simulations are done via Matlab software.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"30 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128943621","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420624
Nasser Sadeghi, M. Azghani
The channel estimation of the muti-user massive MIMO systems is a crucial task which enables us to leverage their high degrees of freedom. Due to the large number of base station antennas and consequently the huge number of channel paths, the massive MIMO channel estimation becomes more challenging. In this paper, we suggest a sparsity-based algorithm to estimate the channels more efficiently. To this end, we would offer a problem modelling to exploit the spatial correlation among different antennas of the BS as well as the inter-user similarity of the channel supports. An iterative thresholding technique has been suggested to approximate the channel matrix. The simulation results confirm that the proposed method has outstanding performance compared to its counterparts.
{"title":"Channel Estimation using Block Sparse Joint Orthogonal Matching Pursuit in Massive MIMO Systems","authors":"Nasser Sadeghi, M. Azghani","doi":"10.1109/CSICC52343.2021.9420624","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420624","url":null,"abstract":"The channel estimation of the muti-user massive MIMO systems is a crucial task which enables us to leverage their high degrees of freedom. Due to the large number of base station antennas and consequently the huge number of channel paths, the massive MIMO channel estimation becomes more challenging. In this paper, we suggest a sparsity-based algorithm to estimate the channels more efficiently. To this end, we would offer a problem modelling to exploit the spatial correlation among different antennas of the BS as well as the inter-user similarity of the channel supports. An iterative thresholding technique has been suggested to approximate the channel matrix. The simulation results confirm that the proposed method has outstanding performance compared to its counterparts.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126050640","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-03-03DOI: 10.1109/CSICC52343.2021.9420599
Sahand Abbasi, Haniyeh Abdi, A. Ahmadi
Since the beginning of the COVID-19 pandemic, many lives are in danger. According to WHO (World Health Organization)’s statements, breathing without a mask is highly dangerous in public and crowded places. Indeed, wearing masks reduces the chance of being infected, and detecting unmasked people is a waste of resources if not performed automatically. AI techniques are used to increase the detection speed of masked and unmasked faces. In this research, a novel dataset and two different methods are proposed to detect masked and unmasked faces in real-time. In the first method, an object detection model is applied to find and classify masked and unmasked faces. In the second method, a YOLO face detector spots faces (whether masked or not), and then the faces are classified into masked and unmasked categories with a novel fast yet effective CNN architecture. By the methods proposed in this paper, the accuracy of 99.5% is achieved on the newly collected dataset.
{"title":"A Face-Mask Detection Approach based on YOLO Applied for a New Collected Dataset","authors":"Sahand Abbasi, Haniyeh Abdi, A. Ahmadi","doi":"10.1109/CSICC52343.2021.9420599","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420599","url":null,"abstract":"Since the beginning of the COVID-19 pandemic, many lives are in danger. According to WHO (World Health Organization)’s statements, breathing without a mask is highly dangerous in public and crowded places. Indeed, wearing masks reduces the chance of being infected, and detecting unmasked people is a waste of resources if not performed automatically. AI techniques are used to increase the detection speed of masked and unmasked faces. In this research, a novel dataset and two different methods are proposed to detect masked and unmasked faces in real-time. In the first method, an object detection model is applied to find and classify masked and unmasked faces. In the second method, a YOLO face detector spots faces (whether masked or not), and then the faces are classified into masked and unmasked categories with a novel fast yet effective CNN architecture. By the methods proposed in this paper, the accuracy of 99.5% is achieved on the newly collected dataset.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114889795","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2021-02-25DOI: 10.1109/CSICC52343.2021.9420605
Nazanin Sabri, Ali Edalat, B. Bahrak
The rapid production of data on the internet and the need to understand how users are feeling from a business and research perspective has prompted the creation of numerous automatic monolingual sentiment detection systems. More recently however, due to the unstructured nature of data on social media, we are observing more instances of multilingual and code-mixed texts. This development in content type has created a new demand for code-mixed sentiment analysis systems. In this study we collect, label and thus create a dataset of Persian-English code-mixed tweets. We then proceed to introduce a model which uses BERT pretrained embeddings as well as translation models to automatically learn the polarity scores of these Tweets. Our model outperforms the baseline models that use Naïve Bayes and Random Forest methods.
{"title":"Sentiment Analysis of Persian-English Code-mixed Texts","authors":"Nazanin Sabri, Ali Edalat, B. Bahrak","doi":"10.1109/CSICC52343.2021.9420605","DOIUrl":"https://doi.org/10.1109/CSICC52343.2021.9420605","url":null,"abstract":"The rapid production of data on the internet and the need to understand how users are feeling from a business and research perspective has prompted the creation of numerous automatic monolingual sentiment detection systems. More recently however, due to the unstructured nature of data on social media, we are observing more instances of multilingual and code-mixed texts. This development in content type has created a new demand for code-mixed sentiment analysis systems. In this study we collect, label and thus create a dataset of Persian-English code-mixed tweets. We then proceed to introduce a model which uses BERT pretrained embeddings as well as translation models to automatically learn the polarity scores of these Tweets. Our model outperforms the baseline models that use Naïve Bayes and Random Forest methods.","PeriodicalId":374593,"journal":{"name":"2021 26th International Computer Conference, Computer Society of Iran (CSICC)","volume":"277 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117111443","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：