联系我们

论文查询结果	返回搜索

论文编号：	13199
作者编号：	2120193003
上传时间：	2022/6/7 13:33:15
中文题目：	基于文本挖掘的政民互动研究 ——以人民网领导留言板为例
英文题目：	Research on Political-civilian Interaction Based on Text Mining ——Taking Message Board for Leaders of Peoples’ Daily Online as an Example
指导老师：	王芳
中文关键字：	网络政民互动;文本挖掘;电子政务;政务信息资源管理
英文关键字：	Internet Government-civilian Interaction;Text Mining;E-government;Government Information Resource Management
中文摘要：	近年来，随着互联网及信息技术的快速发展，网络空间治理的重要性日益凸显，政府治理相关行为逐步转移到网络空间中来，网络成为政务活动重要媒介与平台。同时随着网络新媒体的多样化，公众与政府的沟通渠道逐渐被拓宽，其双方的沟通意愿也逐步增强。公众与政府可以通过政务微博、政务公众号以及网络政民互动平台等多种途径进行政务信息沟通与传递，不仅极大地提高了政民互动的效果与效率，提升了公众参与度，还促进了政府回应机制的优化及转型。然而，在此高效的政民互动过程中，信息沟通的平台上留下了大量的政务大数据，这些海量的非结构化数据给政府信息资源管理工作既带来了机遇，也带来了相应的挑战。一方面，这些政务大数据蕴含着丰富的公众诉求与民意，具有不可估量的参考价值；另一方面，面对如此大量且无序的数据，如何利用相关信息技术去挖掘其潜在价值对于政府推动电子政务高质量发展及完善政民交流机制都至关重要。 本文立足于网络政民互动领域现有的研究成果，将唯一一个全国性网络政民互动平台——人民网领导留言板作为研究对象，选取2021年全年的大连市留言板的政务数据作为研究样本，对公众留言与政府回应展开以下几个方面的深入研究。首先，利用构建的LDA主题模型对公众留言的主题进行挖掘，同时分析出各主题的频次强度时序演化情况；其次，基于构建的情感词词典建立了针对政务文本的情感分析模型，分析出政府回应与公众留言的情感倾向分布情况及各主题下的情感倾向分布情况，进而对二者的相关性与均值差异进行检验；然后，通过构建政府回应方式与效果水平的评价体系，对政府回应从三个指标的角度进行量化评价；最后，采用Kruskal-Wallis非参数检验方法，分析出公众留言的主题和情感对政府回应方式与效果的各指标存在部分显著性差异影响，并提出相关对策与建议。 本文的相关研究成果一方面为提升政府信息资源管理工作及政民互动质量的优化提供了新的手段与思路，另一方面具有一定的泛化能力，在一定程度上可以为其他领域政务文本的挖掘与分析提供一些参考和借鉴。
英文摘要：	In recent years, with the rapid development of the Internet and information technology, the importance of cyberspace governance has become increasingly prominent, and government governance-related behaviors have gradually shifted to cyberspace, and the Internet has become an important medium and platform for government activities. At the same time, with the diversification of new online media, the communication channels between the public and the government have been gradually broadened, and their willingness to communicate has also gradually increased. The public and the government can communicate and transmit government information through various channels such as government affairs microblogs, government affairs official accounts, and online government-civilian interaction platforms, which not only greatly improves the effect and efficiency of government-civilian interaction, but also improves public participation. The optimization and transformation of the government response mechanism. However, in the process of efficient government-civilian interaction, a large amount of government big data has been left on the information communication platform. These massive unstructured data have brought both opportunities and corresponding opportunities for government information resource management. challenge. On the one hand, these government affairs big data contain rich public appeals and opinions, and have immeasurable reference value; The high-quality development of e-government and the improvement of the government-civilian exchange mechanism are crucial. Based on the existing research results in the field of online government-civilian interaction, this paper takes the only national online government-civilian interaction platform - People's Daily Online Leader Message Board as the research object, and selects the government affairs data of Dalian's message board in 2021 as the research object. samples, and conduct in-depth research on the following aspects of public messages and government responses. First, use the constructed LDA topic model to mine the topics of public messages, and analyze the time-series evolution of the frequency intensity of each topic; secondly, based on the constructed sentiment word dictionary, a sentiment analysis model for government texts is established to analyze the government response The distribution of emotional tendencies with public messages and the distribution of emotional tendencies under each theme, and then test the correlation and mean difference between the two; then, by constructing an evaluation system for government response methods and effect levels, the government responses are analyzed from three perspectives. Finally, the Kruskal-Wallis non-parametric test method is used to analyze the influence of the subject and emotion of public messages on the government's response mode and effect of each index, and propose relevant countermeasures and suggestions. On the one hand, the relevant research results of this paper provide new means and ideas for improving the management of government information resources and the optimization of the quality of government-civilian interaction; Mining and analysis provide some references.
查看全文：	预览  下载(下载需要进行登录)