基于LDA主题模型和偏序集的在线商品评论研究

Research on Online Commodity Review Based on LDA Topic Model and Partial Order Set

[目的/意义]从大量在线商品评论中得到有价值的信息,为商家掌握顾客消费需求以及消费者选择商品提供依据。[方法/过程]以LDA主题模型和偏序集理论为基础,构建在线商品评论分析模型,并以天猫商城空调的在线评论文本为例进行实证分析。首先,利用八爪鱼采集器爬取商品评论文本;其次,利用LDA主题模型从在线商品评论文本中提取出影响顾客消费的因素;最后,基于偏序集理论对调查问卷收集的数据进行商品排序。[结果/结论]通过实证分析证明该模型可行有效,不仅能帮助商家掌握顾客消费需求和潜在倾向,也可以使消费者迅速筛选出心仪的商品。

[Purpose/significance]The valuable information obtained from a large number of online commodity reviews can provide a basis for merchants to grasp the consumer’s consumption demands and choice of commodities.[Method/process]Based on LDA topic model and partial ordered set theory,the paper constructs an online commodity review analysis model,and takes the online review text of Tmall air conditioner as an example for empirical analysis.The paper uses octopus collector to crawl the product review text,adopts LDA topic model to extract the factors influencing customer consumption from online product review text,and ranks the commodities by the data collected from the questionnaire based on partial ordered set theory.[Result/conclusion]The empirical analysis proves that the model is feasible and effective,which can not only help merchants to grasp the consumer’s consumption demands and potential tendency,but also enable consumers to quickly screen out the desired products.