企业政策匹配数据

采集国家级、省级、市级、县市区级政府网站发布政策，通过政策智能匹配模型为企 业匹配、推送适用的政策，帮助企业及时知晓惠企政策，同时平台根据企业适用的政 策，提供对应的企业服务，帮助企业进行政策申报，享受政策补贴。1.数据采集：通过网络爬虫技术抓取政策，从政府网站、公告板、新闻网站等自动收集政策数据；2.NLP自然语言处理技术：通过 NLP 自然语言处理技术对已有政策数据进行自动聚类，让计算机通过预先阅读各个类目的政策文档并提取特征，如政策名称、政策文号、发布部门、适用地区等。基于提取的特征，为政策文档生成更细化的标签，如适用行业、政策类别、政策标签等，并自动完成政策的内容归类，再经过人工核实确保政策拆解正确;3.智能匹配：通过构建图神经网络计算引擎对政策和企业进行匹配，匹配条件为企业地址和政策适用地区匹配，企业行业领域和政策适用行业匹配，企业类型和政策适用对象，企业标签和政策标签四项中任意二项条件匹配，则政策为企业的适用政策。

This dataset is collected from policy documents released on national, provincial, municipal, and county-level government websites. An intelligent policy matching model is used to match and push applicable policies to enterprises, helping enterprises timely learn about enterprise-benefiting policies. Meanwhile, the platform provides corresponding enterprise services based on the policies applicable to the enterprise, assisting enterprises in policy declaration and enjoying policy subsidies.
1. Data Collection: Policy documents are scraped via web crawling technology, and policy data is automatically collected from government websites, bulletin boards, news websites, and other platforms.
2. Natural Language Processing (NLP) Technology: Existing policy data is automatically clustered using NLP techniques. The computer pre-reads policy documents under various categories and extracts features such as policy name, document number, issuing department, and applicable region. Based on the extracted features, more refined tags are generated for policy documents, including applicable industry, policy category, and policy tags. Automatic content classification of policies is completed, followed by manual verification to ensure the correct decomposition of policies.
3. Intelligent Matching: A graph neural network computing engine is constructed to match policies and enterprises. The matching criterion is that a policy is deemed applicable to an enterprise if any two of the following four conditions are satisfied: 1) the enterprise's address matches the policy's applicable region; 2) the enterprise's industry field matches the policy's applicable industry; 3) the enterprise's type matches the policy's applicable target; 4) the enterprise's tags match the policy's tags.