基于迭代法的产品评论情感倾向分析
首发时间:2013-10-30
摘要:针对产品评论的情感倾向性判断是情感分析研究领域的重点课题。对产品评论进行情感倾向分析,就是对评论文本表达的态度和情感进行分析,判断是肯定还是否定。本文建立了基于弱监督迭代法的情感倾向分析模型,并将其应用到不同评论领域中进行情感倾向分析。方法只需要少量种子情感词汇作引导,通过迭代计算,不断扩充、完善情感词典,在三个层次上计算单词的情感分数,最终完成对篇章级情感倾向分析工作。实验表明在判定不同领域的评论情感倾向性时,表现出了较好的准确性和适应性。
For information in English, please click here
Sentiment Analysis using an Iterative Method
Abstract:Sentiment analysis is an important research topic in natural language processing. Most work focuses on product review analysis. The goal is to determine whether a review is positive or negative. In this paper, we build an iterative method of sentiment analysis and apply it to determine the polarity of product reviews. It only requires a small amount of seed words as the initial sentiment dictionary. The dictionary is then expanded iteratively. The polarity is determined based on the dictionary at zone level, sentence level, and finally the whole review. The experimental results in various domains show that it achieves high accuracy.
Keywords: sentiment analysis, weakly supervised method, iterative model, product review
论文图表:
引用
No.****
同行评议
勘误表
基于迭代法的产品评论情感倾向分析
评论
全部评论0/1000