不同文本长度的体验型产品在线评论时间序列研究——以电影评论为例

目的/意义]将体验型产品在线评论按照文本长度分为长文本在线评论和短文本在线评论，探究这两类评论的时间和内容特征，为电子商务平台掌握消费者在线评论行为规律和商品需求偏好提供情报依据。方法/过程]利用Python爬虫语言获取电影评论网站中在线评论的相关信息，构造在线评论时间间隔序列，基于人类行为动力学相关构念，探究不同类型在线评论发布行为的时间特征规律；利用文本挖掘方法找出不同类型在线评论的文本内容特征并进行比较分析。结果/结论]以电影评论网站在线评论为数据来源，从时间角度总结出不同类型在线评论行为的时间间隔序列符合幂率分布；从文本内容角度发现不同类型在线评论的文本内容特征既有一定的相似性，也表现出明显的差异。

Study on Time Series of Online Experiential Product Review Based on Text Length: Taking Movie Reviews as an Example

Purpose/significance] According to the text length,the online experiential product review is divided into long text online review and short text online review. Exploring the temporal and content characteristic of these two types of online review provides intelligence basis to e-commerce platform about consumers' online review behavior and product demand preference.Method/process] Python crawler language is employed to collect information of online review in movie review website,and then the paper constructs an online comment interval sequence. Human behavioral dynamics theory is used to find out time characteristic law in different types of online review,and on the other hand,text mining method is used to discover content characteristics in different types of online review. The characteristics are compared and analyzed in the paper.Result/conclusion] Taking the movie review websites' online reviews as the data source, from the time perspective,this paper concludes that time interval sequence obeys to the power-law distribution between different types of online review behavior,and from the text mining perspective,it finds that the content characteristics performance similarities as well as significant differences.