作者: Shaojie Pei , Lumin Zhang , Aiping Li
DOI: 10.1007/978-3-319-11119-3_12
关键词:
摘要: Millions of users share their opinions in microblog every day, which makes sentiment analysis an important and practical issue social networks. In this paper, the problem public analysis, construction emoticon networks model are solved using emoticons. Based on large-scale corpus, FP-growth algorithm combined with semantic similarity is proposed to aggregate similar The based Pointwise Mutual Information. And a orientation framework for both messages non-emoticon presented. Experimental results have shown that our approach works effectively analysis.