关键词: 情感描述值, 情感分类, 流数据

Abstract: Along with the rapid development of Internet information, more and more people join in as information maker and this has made a higher requirement for information processing. It’s important to calculate the sentiment description value of the text for evaluating the polarity of the text. Firstly, this paper preprocesses the text to choose the sentences that may best determine the polarity. Secondly, it calculates the sentiment description value of each sub-sentence. Finally, the total value of the combined results is got. This paper also conducts a comprehensive analysis of the elements that may influence the results, such as the length of the text, syntax structure, etc. The experimental results show that the proposed algorithm gets higher accuracy and higher speed, and has better applicability in the situation of large scale stream data environment.

Key words: sentiment description value, sentiment classification, stream data