政府数据开放平台用户评论情感差异分析

Analysis of Sentiment Differences in User Comments of Government Data Open Platform

根据用户在政府数据开放平台的评论反馈,通过主题分类进行情感分析,明确平台提供服务的用户满意度及存在的问题,为优化开放数据平台的建设提供新的分析思路。利用LDA模型对武汉市政府数据开放网站的用户评论数据进行主题提取,结合深度神经网络进行评论分类,并在此基础上进行情感分析,对不同类型的评论情感差异进行探讨。LDA模型共提取9个分类主题,结合情感分析结果,2个主题的情感趋向是满意状态,7个主题的情感趋向是一般或不满意状态,根据分析结果总结平台服务中存在的不足,并提出相应的优化策略。

According to users' comments and feedback on the open platform of government data, this paper makes sentiment analysis through subject classification, clarifies users' satisfaction with the service provided by the platform and the existing problems, and provides new analysis ideas for optimizing the construction of open data platform. Firstly, the LDA model is used to extract the topic from the user comment data of Wuhan Government Open Data Website, and the deep neural network is used to classify the comment. On this basis, sentiment analysis is carried out to explore the emotional differences of different types of comment. LDA model extracts nine categories of topics. Combining with the results of emotional analysis, the emotional trend of two topics is satisfactory, and the emotional trend of six topics is general or unsatisfactory. Based on the analysis results, the deficiencies in platform services are summarized and the corresponding optimization strategies are put forward.