摘要
大数据精细化能力如何有效嵌入政府治理行为,解决实际治理难题。本文以城市治理中最突出的"群租房"治理难题为切入点,利用大数据爬取与分析技术,探索在现实社会问题中,如何利用大数据手段促进政府治理决策的精细化、科学性。以北京市群租房为研究对象,利用机器学习方法和地理信息数据探索准确定位、识别群租房的有效途径。研究发现以公开群租房数据为训练集,机器学习的分类算法可以从公开的租房信息中有效区分群租房和非群租房,这为解决因隐蔽性而难以有效治理群租房问题提供了新的治理路径,也为政府应对诸多社会治理难题提供新的治理方式,是提高政府治理能力现代化的重要着力点。由于现有披露的数据局限,本文对群租房的识别只能达到一定精度,分析模型的有效性还有待提高。
How to combine the big data technology with government behaviors efficiently to solve complex and practical problems.In this paper,we take the most prominent problem'Group Leasing'as a starting point,using big data mining and machine learning techniques to explore whether or not big data could promote the refinement and scientificity of government decision-making in real social problems.Specifically,with the help of machine learning tools,we explore the algorithms that can effectively conduct geographic positioning of group leasing in Beijing.After comparing different classification algorithms of machine learning,we found that using support vector machine(SVM)to detect group-leasing in rental ads is the most effective approach.Thus,big data technology provides a new path for solving concealed group leasing.Furthermore,it is a new way for the government to respond to many social problems that may improve the modernization of government governance capabilities.
作者
蒋林秀
李泉
Jiang Linxiu;Li Quan
出处
《公共管理与政策评论》
CSSCI
2019年第2期85-96,共12页
Public Administration and Policy Review
关键词
群租房
精准治理
机器学习
Group Leasing
Targeted Governance
Machine Learning