基于LDA主题模型的“双一流”高校图书馆用户评论文本数据挖掘

Text Data Mining of User Reviews in“Double First-class”University Libraries Based on LDA Topic Model

[目的/意义]图书馆用户评论中包含用户对图书馆服务和管理意见的重要信息,对用户评论文本进行深度挖掘,旨在探究用户关心主题及其情感态度、隐含诉求,为高校图书馆建设提供理论依据和数据支持。[方法/过程]以“双一流”高校图书馆为研究对象,收集用户在大众点评网上的评论数据,通过LDA主题建模,得到当前用户评论主要集中在信息资源建设、娱乐休闲服务、馆内设施环境三个维度,进而分别分析三个维度下的关键词共现网络和用户情感态度。[结果/结论]用户对高校图书馆的总体情感态度是积极正向的,但在纸质资源管理、社会化服务、馆内人性化服务等方面表现出负面情绪。

[Purpose/significance]The library user reviews contain important information about users’opinions on library services and management.The in-depth analysis of user review texts aims to explore the topics users are concerned about,their emotional attitudes,and implicit demands,so as to provide theoretical basis and data support for the construction of university libraries.[Method/process]This paper takes the“double first-class”university libraries as the research objects,collects user reviews data on Dianping.com,and through LDA topic modeling,gets that the current user reviews mainly focus on three dimensions:information resource construction,entertainment and leisure services,and library facilities and environment.The keyword co-occurrence network and users’emotional attitude under the three dimensions are then analyzed separately.[Result/conclusion]The overall emotional attitude of users to university libraries presents positive,but they show negative emotions in paper resource management,social services,and humanized services in the library.