用户生成内容的图书主题标签研究——以豆瓣读书用户生成评论为例  

Research on Book Topic Tags Based on UGC:Taking the Social Network Platform Douban Reading as an Example

作　　者：陈婵 CHEN Chan(Library of Party School of the C.P.C Wuhan Municipal Committee,Wuhan 430024,China)

机构地区：[1]中共武汉市委党校图书馆,武汉430024

出　　处：《文献与数据学报》2020年第1期80-88,共9页Journal of Library and Data

摘　　要：[目的/意义]为社交网络平台管理者提供一种分析和优化主题标签的方法,以帮助用户在社交网络平台上准确地获取图书的相关信息,满足用户个性化检索的需求。[方法/过程]本文提出了一种基于用户生成内容(UGC)的主题分析方法。以社交平台“豆瓣读书”为例,选取《平凡的世界》和《围城》两本经典图书,首先爬取用户对该书的评价数据,然后对数据进行清洗,基于Latent Dirichlet Allocation(LDA)主题分析方法对数据进行分析,以获取图书的相关主题标签。[结果/结论]通过基于用户生成内容对图书的主题进行分析,一方面完善了图书的标签,提高用户对书籍的查准率,另一方面用户生成内容中具有鲜明的主题性和情感倾向,因此豆瓣读书制作标签时可以考虑增加情感类词,提高社交网络平台的个性化推荐功能。[Purpose/meaning]This paper provides a way to analyze and optimize book topic tags for social network platform manager,in order to help users to obtain relevant information of books accurately on the social network platform,and meet the needs of users for personalized retrieve.[Method/process]This paper proposes a topic analysis method based on user generated content(UGC),takes the social platform Douban Reading as an example and selects two classic books,Ordinary World and The Besieged City.The process includes:(1)climbing the UGC data of the book;(2)cleaning the data;(3)analyzing the data based on the Latent Dirichlet Allocation(LDA)theme analysis method to obtain the relevant topic labels of books.[Result/conclusion]By getting and analyzing the topic tags of the books based on the UGC,on the one hand,the tags of the books are improved,and the precision of the user's book information retrieve is improved.On the other hand,the UGC has a distinct theme and emotional tendency.Therefore,when making topic tags,Douban reading can consider adding emotional words to improve the personalized recommendation function of the social network platform.

关 键 词：UGC 图书标签 LDA 

分 类 号：G250[文化科学—图书馆学]