The effect of aggregation methods on sentiment classification in Persian reviews

被引:16
作者
Basiri M.E. [1 ]
Kabiri A. [1 ]
Abdar M. [2 ]
Mashwani W.K. [3 ]
Yen N.Y. [4 ]
Hung J.C. [5 ]
机构
[1] Department of Computer Engineering, Shahrekord University, Shahrekord
[2] Department of Computer Science, University of Quebec in Montreal, Montreal, QC
[3] Department of Mathematics, Kohat University of Science and Technology, Kohat
[4] School of Computer Science and Engineering, The University of Aizu, Aizu-Wakamatsu
[5] Department of Computer Science and Information Engineering, National Taichung University of Science and Technology, Taichung
关键词
aggregation methods; lexicon-based method; Persian language processing; Sentiment analysis;
D O I
10.1080/17517575.2019.1669829
中图分类号
学科分类号
摘要
One of the most essential parts of every sentiment analysis application is the aggregation mechanism used to combine results obtained from a lower granularity level into an overall result. In this paper, the effects of the sentiment lexicon, aggregation level, and aggregation method on the sentiment polarity and rating classification of Persian reviews are investigated. To this aim, a new sentiment aggregation method based on the cross-ratio operator is proposed. The results on four Persian review data sets show that the review-level aggregation can improve rating classification, although this approach does not have a positive impact on polarity classification. © 2019 Informa UK Limited, trading as Taylor & Francis Group.
引用
收藏
页码:1394 / 1421
页数:27
相关论文
共 66 条
[1]  
Abbasi A., Chen H., Salem A., Sentiment Analysis in Multiple Languages, ACM Transactions on Information Systems, 26, 3, pp. 1-34, (2008)
[2]  
Alimardaniaghaie S.A., Opinion Mining in Persian Language Using Supervised Algorithms, Journal of Information Systems and Telecommunication, 3, (2015)
[3]  
Appel O., Et al., A Hybrid Approach to the Sentiment Analysis Problem at the Sentence Level, Knowledge-Based Systems, 108, pp. 110-124, (2016)
[4]  
Appel O., Et al., Cross-ratio Uninorms as an Effective Aggregation Mechanism in Sentiment Analysis, Knowledge-Based Systems, 124, pp. 16-22, (2017)
[5]  
Appel O., Et al., IOWA & Cross-ratio Uninorm Operators as Aggregation Tools in Sentiment Analysis and Ensemble Methods, 2017 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), pp. 1-6, (2017)
[6]  
NLPTools, (2016)
[7]  
Asgarian E., Kahani M., Sharifi S., The Impact of Sentiment Features on the Sentiment Polarity Classification in Persian Reviews, Cognitive Computation, 10, 1, pp. 117-135, (2018)
[8]  
Bagherim A.S., Persian Sentiment Analyzer: A Framework Based on A Novel Feature Selection Method.” International, Journal of Artificial Intelligence, 12, 2, pp. 115-129, (2014)
[9]  
Bagheri A.M., Saraeede Jong F., Sentiment Classification in Persian: Introducing a Mutual Information-based Method for Feature Selection, 2013 21St Iranian Conference on Electrical Engineering (ICEE), pp. 1-6, (2013)
[10]  
Basirikabiri M.E.A., Sentence-level Sentiment Analysis in Persian, 2017 3Rd International Conference on Pattern Recognition and Image Analysis (IPRIA)., pp. 84-89, (2017)