{"title":"The Repository of Web Document Summarization using Social Information","authors":"Minh-Tien Nguyen, Van-Hau Nguyen, Duc-Vu Tran","doi":"10.1109/KSE.2019.8919378","DOIUrl":null,"url":null,"abstract":"Summarization using social information is a task which extracts summary sentences and relevant user posts of a Web document by integrating its relevant social information. Prior studies introduced several strong models for this task; however, there are gaps from papers to the reproduction of such models. This paper leverages the gaps by investigating summa-rization algorithms to facilitate next studies. The investigation was conducted by implementing traditional and state-of-the-art methods, from unsupervised to supervised learning fashion. We used three datasets in English and Vietnamese to confirm the efficiency of the methods. Experimental results indicate that sophisticated models obtain improvements of ROUGE-scores compared to the basic ones, which do not use social information. However, in some cases, simple methods comparably perform state-of-the-art methods, suggesting that the performance of summarization methods can be still improved.","PeriodicalId":439841,"journal":{"name":"2019 11th International Conference on Knowledge and Systems Engineering (KSE)","volume":"221 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2019-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2019 11th International Conference on Knowledge and Systems Engineering (KSE)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/KSE.2019.8919378","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Summarization using social information is a task which extracts summary sentences and relevant user posts of a Web document by integrating its relevant social information. Prior studies introduced several strong models for this task; however, there are gaps from papers to the reproduction of such models. This paper leverages the gaps by investigating summa-rization algorithms to facilitate next studies. The investigation was conducted by implementing traditional and state-of-the-art methods, from unsupervised to supervised learning fashion. We used three datasets in English and Vietnamese to confirm the efficiency of the methods. Experimental results indicate that sophisticated models obtain improvements of ROUGE-scores compared to the basic ones, which do not use social information. However, in some cases, simple methods comparably perform state-of-the-art methods, suggesting that the performance of summarization methods can be still improved.