The paradigm of cloud robotics points out a direction for the future development of robots. By deploying robotic applications in the cloud, the workload and cost of local robots are greatly reduced. The rise of microservices and cloud-native technology provides conveniences and guarantees for the development and deployment of cloud applications. This paper proposes a cloud robotic application platform design based on microservices. With the help of Robot Operating System (ROS), we can use the existing rich and diverse robot software packages and deploy them in the cloud without extra modifications. Through the microservices architecture and container technology, robotic applications can be further decoupled in the cloud. That improves the flexibility and compatibility of the platform and embodies the core idea of microservices. In the end, we present a demonstration to cooperate with a simulated robot to complete the simultaneous localization and mapping (SLAM) task, which verifies the feasibility of our design.
{"title":"A Cloud Robotic Application Platform Design Based on the Microservices Architecture","authors":"Binhuai Xu, Jing Bian","doi":"10.1145/3437802.3437805","DOIUrl":"https://doi.org/10.1145/3437802.3437805","url":null,"abstract":"The paradigm of cloud robotics points out a direction for the future development of robots. By deploying robotic applications in the cloud, the workload and cost of local robots are greatly reduced. The rise of microservices and cloud-native technology provides conveniences and guarantees for the development and deployment of cloud applications. This paper proposes a cloud robotic application platform design based on microservices. With the help of Robot Operating System (ROS), we can use the existing rich and diverse robot software packages and deploy them in the cloud without extra modifications. Through the microservices architecture and container technology, robotic applications can be further decoupled in the cloud. That improves the flexibility and compatibility of the platform and embodies the core idea of microservices. In the end, we present a demonstration to cooperate with a simulated robot to complete the simultaneous localization and mapping (SLAM) task, which verifies the feasibility of our design.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"112 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122888565","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}
This survey talks about Automatic Text Summarization. Information explosion, the problem caused by the rapid growth of the internet, increased more and more necessity of powerful summarizers. This article briefly reviews different methods and evaluation metrics. The main attention is on the applications of the latest trends, neural network-based, and pre-trained transformer language models. Pre-trained language models now are ruling the NLP field, as one of the main down-stream tasks, Automatic Text Summarization is quite an interdisciplinary task and requires more advanced techniques. But there is a limitation of input and context length results in that the whole article cannot be encoded completely. Motivated by the application of recurrent mechanism in Transformer-XL, we build an abstractive summarizer for long text and evaluate how well it performs on dataset CNN/Daily Mail. The model is under general sequence to sequence structure with a recurrent encoder and stacked Transformer decoder. The obtained ROUGE scores tell that the performance is good as expected.
{"title":"Survey on Automatic Text Summarization and Transformer Models Applicability","authors":"Guan Wang, I. Smetannikov, T. Man","doi":"10.1145/3437802.3437832","DOIUrl":"https://doi.org/10.1145/3437802.3437832","url":null,"abstract":"This survey talks about Automatic Text Summarization. Information explosion, the problem caused by the rapid growth of the internet, increased more and more necessity of powerful summarizers. This article briefly reviews different methods and evaluation metrics. The main attention is on the applications of the latest trends, neural network-based, and pre-trained transformer language models. Pre-trained language models now are ruling the NLP field, as one of the main down-stream tasks, Automatic Text Summarization is quite an interdisciplinary task and requires more advanced techniques. But there is a limitation of input and context length results in that the whole article cannot be encoded completely. Motivated by the application of recurrent mechanism in Transformer-XL, we build an abstractive summarizer for long text and evaluate how well it performs on dataset CNN/Daily Mail. The model is under general sequence to sequence structure with a recurrent encoder and stacked Transformer decoder. The obtained ROUGE scores tell that the performance is good as expected.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121258321","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}
Jiaxuan Fei, Kai Chen, Qigui Yao, Qian Guo, Xiangqun Wang
Power Internet of Things is the application of IoT in smart power grid. Once attacked, it will cause huge losses. Therefore, it is necessary to conduct a security assessment to take defensive measures. However, the traditional vulnerability assessment methods of the power Internet of things mostly focus on the security of the system itself, without considering the impact on business economy and efficiency. This paper proposes a security vulnerability assessment method of power Internet of Things integrating business security. This method first analyzes the security risks faced by the power Internet of Things, and establishes its attack tree model. Then, each leaf node is rated from the three safety features, which are weighted by evaluation and calculation, and the activation probability of each leaf node is calculated. After that, considering the blind attack factor, the activation probability of all nodes in the model is calculated. Finally, the vulnerability of the system and the vulnerability sensitivity of each leaf node are obtained. According to the vulnerability sensitivity, measures are taken to protect the weak links of the system. The effectiveness of the proposed method is verified by experiments on SCADA (supervisory control and data acquisition) system in the power Internet of things.
{"title":"Security Vulnerability Assessment of Power IoT based on Business Security","authors":"Jiaxuan Fei, Kai Chen, Qigui Yao, Qian Guo, Xiangqun Wang","doi":"10.1145/3437802.3437825","DOIUrl":"https://doi.org/10.1145/3437802.3437825","url":null,"abstract":"Power Internet of Things is the application of IoT in smart power grid. Once attacked, it will cause huge losses. Therefore, it is necessary to conduct a security assessment to take defensive measures. However, the traditional vulnerability assessment methods of the power Internet of things mostly focus on the security of the system itself, without considering the impact on business economy and efficiency. This paper proposes a security vulnerability assessment method of power Internet of Things integrating business security. This method first analyzes the security risks faced by the power Internet of Things, and establishes its attack tree model. Then, each leaf node is rated from the three safety features, which are weighted by evaluation and calculation, and the activation probability of each leaf node is calculated. After that, considering the blind attack factor, the activation probability of all nodes in the model is calculated. Finally, the vulnerability of the system and the vulnerability sensitivity of each leaf node are obtained. According to the vulnerability sensitivity, measures are taken to protect the weak links of the system. The effectiveness of the proposed method is verified by experiments on SCADA (supervisory control and data acquisition) system in the power Internet of things.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128881488","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}
V. Kazakovtsev, Svyatoslav Oreshin, A. Serdyukov, Egor Krasheninnikov, S. Muravyov, Albert Bezvinnyi, Alexander Panfilov, Igor Glukhov, Y. Kaliberda, Daniil Masalskiy, Timofey Podolenchuk, Maksim Khlopotov
This article proposes a recommendation system for choosing an academic supervisor, based on an assessment of the similarity of student interests and the scientific achievements of the possible mentor from the university faculty. We used a new approach to calculate similarity with no creating co-authorship networks but using Scopus quality metrics. Each scientist is presented as a combination of his achievements in each field of science. As a normalization method, we used the cumulative distribution function of the logarithm of the weighted impacts of professors in the field. We compared different similarity measures and performed clustering to assess their adequacy and thus assess the quality of the system due to the impossibility of comparing the received recommendations with the data of the past years.
{"title":"Recommender system for an academic supervisor with a matrix normalization approach","authors":"V. Kazakovtsev, Svyatoslav Oreshin, A. Serdyukov, Egor Krasheninnikov, S. Muravyov, Albert Bezvinnyi, Alexander Panfilov, Igor Glukhov, Y. Kaliberda, Daniil Masalskiy, Timofey Podolenchuk, Maksim Khlopotov","doi":"10.1145/3437802.3437817","DOIUrl":"https://doi.org/10.1145/3437802.3437817","url":null,"abstract":"This article proposes a recommendation system for choosing an academic supervisor, based on an assessment of the similarity of student interests and the scientific achievements of the possible mentor from the university faculty. We used a new approach to calculate similarity with no creating co-authorship networks but using Scopus quality metrics. Each scientist is presented as a combination of his achievements in each field of science. As a normalization method, we used the cumulative distribution function of the logarithm of the weighted impacts of professors in the field. We compared different similarity measures and performed clustering to assess their adequacy and thus assess the quality of the system due to the impossibility of comparing the received recommendations with the data of the past years.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127844736","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}
The working status of wind turbine can be obtained and fault warning can be given accurately, if the data information mining is efficient. However the existing SCADA data monitoring methods do not take the history and trend into account. A data information mining method based on LSSVR for wind turbine SCADA data is presented in this paper. First, LSSVR model of wind turbine with output power as output and other 30 parameters as input is built by using the SCADA data of wind turbine normal condition. Then, using the LSSVR model, the residual of output power prediction and actual value is obtained. At last, by analyzing the current information, historical information and trend information mined from the residual, wind turbine working status is concluded and early warning is given if necessary. Through cases of both chronic fault and acute fault, the accuracy and effectiveness of the proposed method is verified which means the maintenance cost of WT could be reduced by using the proposed method.
{"title":"Wind Turbine Health Information Mining Based on SCADA Data","authors":"Zhengnan Hou, Xiaoxiao Lv, Shengxian Zhuang","doi":"10.1145/3437802.3437815","DOIUrl":"https://doi.org/10.1145/3437802.3437815","url":null,"abstract":"The working status of wind turbine can be obtained and fault warning can be given accurately, if the data information mining is efficient. However the existing SCADA data monitoring methods do not take the history and trend into account. A data information mining method based on LSSVR for wind turbine SCADA data is presented in this paper. First, LSSVR model of wind turbine with output power as output and other 30 parameters as input is built by using the SCADA data of wind turbine normal condition. Then, using the LSSVR model, the residual of output power prediction and actual value is obtained. At last, by analyzing the current information, historical information and trend information mined from the residual, wind turbine working status is concluded and early warning is given if necessary. Through cases of both chronic fault and acute fault, the accuracy and effectiveness of the proposed method is verified which means the maintenance cost of WT could be reduced by using the proposed method.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124118680","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}
Heng Cao, Chundong Wang, Long Huang, Xiaochun Cheng, Haoran Fu
Botnets have long relied on the Domain Generation Algorithm (DGA) to survive to this day. The detection rate of the DGA detection methods based on machine learning is already high. However, the models trained by the existing data sets sometimes are blind to new variant domains.To mitigate such problem, a method based on generation adversarial networks (GAN) called DnGAN is proposed to generate adversarial DGA examples in this paper. Experiment results show that the adversarial examples can effectively escape the detection of multiple detectors. And by using these adversarial examples as training data can effectively enhance the ability of the detector to identify DGA families that have not been seen before.
{"title":"Adversarial DGA Domain Examples Generation and Detection","authors":"Heng Cao, Chundong Wang, Long Huang, Xiaochun Cheng, Haoran Fu","doi":"10.1145/3437802.3437836","DOIUrl":"https://doi.org/10.1145/3437802.3437836","url":null,"abstract":"Botnets have long relied on the Domain Generation Algorithm (DGA) to survive to this day. The detection rate of the DGA detection methods based on machine learning is already high. However, the models trained by the existing data sets sometimes are blind to new variant domains.To mitigate such problem, a method based on generation adversarial networks (GAN) called DnGAN is proposed to generate adversarial DGA examples in this paper. Experiment results show that the adversarial examples can effectively escape the detection of multiple detectors. And by using these adversarial examples as training data can effectively enhance the ability of the detector to identify DGA families that have not been seen before.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114527144","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}
Svyatoslav Oreshin, A. Filchenkov, Polina Petrusha, Egor Krasheninnikov, Alexander Panfilov, Igor Glukhov, Y. Kaliberda, Daniil Masalskiy, A. Serdyukov, V. Kazakovtsev, Maksim Khlopotov, Timofey Podolenchuk, I. Smetannikov, D. Kozlova
This research is dedicated to the problem of transforming ”linear” educational systems of higher education institutions into a new paradigm of person-centered, blended and individual education. This paper investigates role, application, and challenges of applying AI to predict the academic performance traditional of students: dropouts, GPA, publication activity and other indicators to decrease dropouts and make the learning process more personalized and adaptive. In the first part, we overview the process of data mining using internal university’s resources (LMS and other systems) and open source data from students’ social networks. Such an aggregation allows describing each student by socio-demographic and psychometric features. Further, we demonstrate how we can dynamically monitor students’ activities during the learning process to supplement the resulting features. In the second part of our research, we propose various static and dynamic targets for predictive models and demonstrate the results of predictions and comparisons of several predictive models. The research is based on the information on data processing of more than 20000 students in 2013-2019.
{"title":"Implementing a Machine Learning Approach to Predicting Students’ Academic Outcomes","authors":"Svyatoslav Oreshin, A. Filchenkov, Polina Petrusha, Egor Krasheninnikov, Alexander Panfilov, Igor Glukhov, Y. Kaliberda, Daniil Masalskiy, A. Serdyukov, V. Kazakovtsev, Maksim Khlopotov, Timofey Podolenchuk, I. Smetannikov, D. Kozlova","doi":"10.1145/3437802.3437816","DOIUrl":"https://doi.org/10.1145/3437802.3437816","url":null,"abstract":"This research is dedicated to the problem of transforming ”linear” educational systems of higher education institutions into a new paradigm of person-centered, blended and individual education. This paper investigates role, application, and challenges of applying AI to predict the academic performance traditional of students: dropouts, GPA, publication activity and other indicators to decrease dropouts and make the learning process more personalized and adaptive. In the first part, we overview the process of data mining using internal university’s resources (LMS and other systems) and open source data from students’ social networks. Such an aggregation allows describing each student by socio-demographic and psychometric features. Further, we demonstrate how we can dynamically monitor students’ activities during the learning process to supplement the resulting features. In the second part of our research, we propose various static and dynamic targets for predictive models and demonstrate the results of predictions and comparisons of several predictive models. The research is based on the information on data processing of more than 20000 students in 2013-2019.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"81 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130565967","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}
As for the modern network traffic circulation which has been realized by the Internet, it is one thing to discuss the problem asking how to detect where the network traffic failure has occurred, and quite another to discuss the problem asking how to predict and how to estimate the frequency indicating numerically how often the network traffic failure occurs, because the former problem, which is called the network traffic failure detection problem, and the latter problem, which is called the network traffic failure estimation problem, are investigated with the network skills based on the statistical methods and the network skills based on the probabilistic methods, respectively. Moreover, since it is one thing to locate the network traffic failure on the network segments and quite another to predict them beforehand, it is important for us to apply not only statistical methods but also probabilistic ones for the solutions to these problems.
{"title":"Numerical Estimation of Network Traffic Failure Based on Probabilistic Approximation Methods: To what extent the network traffic failure can be predicted?","authors":"Shigeo Akashi, Yao Tong","doi":"10.1145/3437802.3437838","DOIUrl":"https://doi.org/10.1145/3437802.3437838","url":null,"abstract":"As for the modern network traffic circulation which has been realized by the Internet, it is one thing to discuss the problem asking how to detect where the network traffic failure has occurred, and quite another to discuss the problem asking how to predict and how to estimate the frequency indicating numerically how often the network traffic failure occurs, because the former problem, which is called the network traffic failure detection problem, and the latter problem, which is called the network traffic failure estimation problem, are investigated with the network skills based on the statistical methods and the network skills based on the probabilistic methods, respectively. Moreover, since it is one thing to locate the network traffic failure on the network segments and quite another to predict them beforehand, it is important for us to apply not only statistical methods but also probabilistic ones for the solutions to these problems.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114574270","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}
To improve the classification accuracy of complaint work order text data, a deep learning-based classification method is designed. The word vector of this paper uses word2vec. Although word2vec represents the semantic richness of the words, it ignores the semantic information of the local words of the sentence. The word vector using a combination of n-gram and word2vec is both semantically rich and takes into account the local word order. In terms of the classification model, a combination of attention and CNN is used to consider both global and local features. After several sets of comparative experiments, the proposed algorithm for text classification on a company's complaint text effectively improves the accuracy rate. The accuracy rate is better than other algorithms reaching more than 90%.
{"title":"An improved text classification method based on convolutional neural networks","authors":"Yan Yan, Wenya Li, Guanhua Chen, Wei Liu","doi":"10.1145/3437802.3437833","DOIUrl":"https://doi.org/10.1145/3437802.3437833","url":null,"abstract":"To improve the classification accuracy of complaint work order text data, a deep learning-based classification method is designed. The word vector of this paper uses word2vec. Although word2vec represents the semantic richness of the words, it ignores the semantic information of the local words of the sentence. The word vector using a combination of n-gram and word2vec is both semantically rich and takes into account the local word order. In terms of the classification model, a combination of attention and CNN is used to consider both global and local features. After several sets of comparative experiments, the proposed algorithm for text classification on a company's complaint text effectively improves the accuracy rate. The accuracy rate is better than other algorithms reaching more than 90%.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117313873","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}
Wei Liu, Renze Xiong, Ning N. Cheng, Yiming Y. Sun
To accurately express the fuzzy relation between word features and texts, and fuzzy relation between word features and categories respectively. A text classification method is proposed based on Fuzzy Relation and Feature Distribution Variance (FRFDV). This method firstly performs feature reduction and category feature word extraction according to the distribution of features in inter-category and intra-category. Then the method defines the word feature set, test text set and category set as fuzzy sets. Next, each text and category are represented respectively by defining the membership function of the word feature set to the test text set and the category set. When using word feature sets to represent categories, pay attention to the membership degree of features to categories and their distribution between categories; when using feature sets to represent test texts, give categorical feature words and non-categorical feature words with different weights. Finally, the fuzzy set correlation formula is used to calculate the correlation between the text and each category, and the category with the largest correlation is the category of the text. Comparing with the XGBOOST [Fang, 2020, Gong and Wang, 2018] algorithm and SVM algorithm, it is proved that the text classification method based on FRFDV is feasible. The accuracy of the results is higher by 2 % and 4 % respectively.
准确表达词特征与文本的模糊关系、词特征与类别的模糊关系。提出一种基于模糊关系和特征分布方差(FRFDV)的文本分类方法。该方法首先根据特征在类别间和类别内的分布进行特征约简和类别特征词提取。然后将词特征集、测试文本集和类别集定义为模糊集。接下来，通过定义单词特征集对测试文本集和类别集的隶属函数来分别表示每个文本和类别。在使用词特征集表示类别时，要注意特征与类别的隶属度及其在类别之间的分布;在使用特征集表示测试文本时，给出不同权重的分类特征词和非分类特征词。最后，利用模糊集关联公式计算文本与各类别之间的关联，关联度最大的类别即为该文本所属的类别。对比XGBOOST [Fang, 2020, Gong and Wang, 2018]算法和SVM算法，证明了基于FRFDV的文本分类方法是可行的。结果的准确度分别提高了2%和4%。
{"title":"Text Classification Method with Combination of Fuzzy Relation and Feature Distribution Variance","authors":"Wei Liu, Renze Xiong, Ning N. Cheng, Yiming Y. Sun","doi":"10.1145/3437802.3437829","DOIUrl":"https://doi.org/10.1145/3437802.3437829","url":null,"abstract":"To accurately express the fuzzy relation between word features and texts, and fuzzy relation between word features and categories respectively. A text classification method is proposed based on Fuzzy Relation and Feature Distribution Variance (FRFDV). This method firstly performs feature reduction and category feature word extraction according to the distribution of features in inter-category and intra-category. Then the method defines the word feature set, test text set and category set as fuzzy sets. Next, each text and category are represented respectively by defining the membership function of the word feature set to the test text set and the category set. When using word feature sets to represent categories, pay attention to the membership degree of features to categories and their distribution between categories; when using feature sets to represent test texts, give categorical feature words and non-categorical feature words with different weights. Finally, the fuzzy set correlation formula is used to calculate the correlation between the text and each category, and the category with the largest correlation is the category of the text. Comparing with the XGBOOST [Fang, 2020, Gong and Wang, 2018] algorithm and SVM algorithm, it is proved that the text classification method based on FRFDV is feasible. The accuracy of the results is higher by 2 % and 4 % respectively.","PeriodicalId":429866,"journal":{"name":"Proceedings of the 2020 1st International Conference on Control, Robotics and Intelligent System","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120955800","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：