关键词: 情感词汇本体, 主观性句子, 3-POS 模型, 倾向性计算

Abstract: How to effectively extract the subjective information of a sentence, calculate the orientation of the subjective sentence, and analyze the fine-grained sentiment orientation, has become a hot research topic. A kind of analytical method was proposed based on sentiment words ontology, and rules were created through analyzing the semantic content of sentences. The 3-POS model was put forward to identify the subjective sentences, and the orientation value of a subjective sentence was calculated according to emotional vocabulary ontology and rules. The experimental results show that the accuracy of distinguishing and calculating the orientation of subjective sentences is 81.02 percent by using the proposed model.

Key words: sentiment words ontology, subjective sentence, 3-POS model, orientation computation