面向共现关系和情感分析的品牌竞争优势研究

Study on Brand Competitive Advantage Based on Concurrence Relation and Sentiment Analysis

将共词分析应用于产品评论挖掘,实现细粒度的情感分析,识别不同品牌的竞争优势。选取显示器产品评论作为实验语料,使用主题模型对文本降维并提取深层语义形成主题词表,对主题词进行筛选并构建共词网络,分析产品特征和情感词的关联强度。实验结果表明,不同品牌的共词网络中核心特征及情感极性各有不同,共词网络能够识别不同品牌的竞争优势和差异定位。

This paper applies co-occurrence analysis to product review mining,aiming at achieving feature-level sentiment analysis and identifying competitive advantage of different brands. This research collects product reviews about display device as the experimental corpus,and uses LDA model to reduce the dimension of the text and extract the deep semantics to form the subject words list. Through screening subject words that LDA model provides,this paper builds a cooccurrence network to analyze the correlation strength between product characteristics and sentiment words. The experimental results show that the core characteristics and sentiment polarity of co-occurrence network of different brands are different, concurrence analysis can identify the competitive advantages and differential positioning of different brands.