A Target-Dependent Sentiment Analysis Method for Micro-blog Streams
Sentiment analysis technique is useful for companies to analyze customers opinion about products and to find potential customers.Most of the target-dependent sentiment analysis methods can not get acceptable accuracy.Recently some new sentiment analysis methods using Recursive Neural Networks (RNN) are promissing but they are not target-dependent.In this paper we propose a target-dependent sentiment analysis method for micro-blog streams based on RNN.We use cluster-based data partitioning to get higher accuracy with limited labeled samples.A tree pruning method is proposed to remove irrelevant parts from the syntax tree.The original recursive neural network model is extended to support target-dependent sentiment analysis better.Experimental results on two corpuses with different targets show that the performance of our method is better than previous methods.
Yongheng Wang Hui Gao Shaofeng Geng
College of Information Science and Engineering,Hunan University,Changsha 410082,China
国际会议
International Asia-Pacific Web Conference(第18届国际亚太互联网大会)
苏州
英文
30-42
2016-09-23（万方平台首次上网日期，不代表论文的发表时间）