Applying multi-label techniques in emotion identification of short texts

被引:33
作者
Almeida, Alex M. G. [1 ]
Cerri, Ricardo [2 ]
Paraiso, Emerson Cabrera [3 ]
Mantovani, Rafael Gomes [4 ]
Barbon Junior, Sylvio [1 ]
机构
[1] State Univ Londrina UEL, Dept Comp Sci, BR-86057970 Londrina, Brazil
[2] Fed Univ Sao Carlos UFSCar, Dept Comp Sci, BR-13565905 Sao Carlos, SP, Brazil
[3] Pontificia Univ Catolica Parana PUCPR, Grad Program Informat, BR-80215901 Curitiba, Parana, Brazil
[4] Univ Sao Paulo, Inst Math & Comp Sci ICMC, BR-13566590 Sao Carlos, SP, Brazil
关键词
Sentiment analysis; Back propagation multi-label learning; Multi-label k-Nearest Neighbor; Hierarchy of multi-label classifier; Algorithm adaptation; Problem transformation; CLASSIFICATION; CLASSIFIERS; NETWORKS;
D O I
10.1016/j.neucom.2018.08.053
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Sentiment Analysis is an emerging research field traditionally applied to classify opinions, sentiments and emotions towards polarity and subjectivity expressed in text. An important characteristic to automatic emotion analysis is the standpoint, in which we can look at an opinion from two perspectives, the opinion holder (author) who express an opinion, and the reader who reads and perceives the opinion. From the reader's standpoint, the interpretations of the text can be multiple and depend on the personal background. The multiple standpoints cognition, in which readers can look at the same sentence, is an interesting scenario to use the multi-label classification paradigm in the Sentiment Analysis domain. This methodology is able to handle different target sentiments simultaneously in the same text, by also taking advantage of the relations between them. We applied different approaches such as algorithm adaptation, problem transformation and ensemble methods in order to explore the wide range of multi-label solutions. The experiments were conducted on 10,080 news sentences from two different real datasets. Experimental results showed that the Ensemble Classifier Chain overcame the other algorithms, average F-measure of 64.89% using emotion strength features, when considering six emotions and neutral sentiment. (C) 2018 Elsevier B.V. All rights reserved.
引用
收藏
页码:35 / 46
页数:12
相关论文
共 51 条
[1]  
Almeida AMG, 2016, PROCEEDINGS OF 2016 5TH BRAZILIAN CONFERENCE ON INTELLIGENT SYSTEMS (BRACIS 2016), P486, DOI [10.1109/BRACIS.2016.92, 10.1109/BRACIS.2016.093]
[2]  
[Anonymous], 2001, Emotions in social psychology: Essential readings
[3]  
[Anonymous], 2012, EACL 2012
[4]  
[Anonymous], 2014, Evaluating Learning Algorithms A Classification Perspective, DOI DOI 10.1017/CBO9780511921803
[5]  
[Anonymous], INT FLOR ART INT RES
[6]  
Balahur A, 2010, LREC 2010 - SEVENTH INTERNATIONAL CONFERENCE ON LANGUAGE RESOURCES AND EVALUATION, P2216
[7]  
Bautin M., 2008, P ICWSM, P1
[8]  
Bhowmick P.K., 2009, Comput. Inf. Sci, V2, P64
[9]  
Blockeel H., 1998, Machine Learning. Proceedings of the Fifteenth International Conference (ICML'98), P55
[10]  
Cambria Erik, 2012, Cognitive Behavioural Systems (COST 2012). International Training School. Revised Selected Papers, P144, DOI 10.1007/978-3-642-34584-5_11