Abstract | ||
---|---|---|
Short text is the most commonly used form of expression in the network. As short texts like microblog do not provide sufficient word occurrences, sentiment classification methods that use traditional approaches have limitations. In this paper, we propose a short text sentiment classification model called FECEM base on short text entropy optimization method. This method first selects sentiment features based on expectation cross entropy, and then fuzzy sets is used to correct the degree of the comment words. Experiments show that our method is more efficient than the SVM+Maximum Entropy and SVM+ chi-square methods, and this new method is robust across different types of short text. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1109/WISA.2015.22 | IEEE WISA |
Keywords | Field | DocType |
expectation cross entropy, short text, sentiment classification, fuzzy sets, emotional feature | Cross entropy,Data mining,Social media,Pattern recognition,Sentiment analysis,Microblogging,Support vector machine,Feature extraction,Fuzzy set,Mutual information,Artificial intelligence,Principle of maximum entropy | Conference |
ISBN | Citations | PageRank |
978-1-4673-9371-3 | 2 | 0.38 |
References | Authors | |
3 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tao Jiang | 1 | 2 | 0.38 |
Bin Yuan | 2 | 10 | 6.32 |
Jing Jiang | 3 | 2 | 0.38 |
Hongzhi Yu | 4 | 4 | 5.18 |