Research On Collection And Organization Of Public Opinion Information For Government Decision Support

In recent years, all kinds of social public opinion events happened frequently, especially with the coming of big data, social public opinion information also highlights big data characteristics of big volume, fast velocity, variety, low value and veracity, and it has become the important basis and data sources for government decision-making. Due to the important application value and practical significance of social public opinion research, more and more scholars has begun to carry out the research of its collection, processing, organization, analysis, early warning and the government decision-making support. To the social public opinion information collection and organization, it is the premise and foundation of analysis. However, it has the characteristics of multi-source, multi-structure and low value, which has brought many problems and challenges for effectively collecting and organizing, also for serving government decision support. Research of this paper is to solve the related problems to support for government decision-making, including social public opinion information collection lacking of target, organization lacking of semantics, public opinion case database describing too large size, and so on.This study focuses on social public opinion information collection and organization based on the actual needs of the government decision makers. Study also builds social public opinion case knowledge base and designs the collection and organization system prototype. Specific perspective, in the theory and method research part, this study integrated using theories and methods including semantic association, literature metrology, ontology, natural language processing, text mining and so on. Firstly, it establishes the public opinion information collection system framework for government decision support based on user demands and the government decision-making research analysis. Secondly, it completes the semantic feature extraction, which achieves the fine-grained description of the social public opinion, through a series of methods and steps, such as Chinese word segmentation, part-of-speech tagging, named entity recognition, emotional extraction and so on. Thirdly, it constructs the POIOM model applicable to describe the social public opinion information according to the seven steps and referring for ABC event ontology model, and selects the OWL ontology description language and ontology construction tools Protege to design it. What’s more, it proposes social public opinion case knowledge base design on the analysis of the existing case base. The case knowledge base builds association according to POIOM ontology and constitutes an organic whole. In the empirical research part, this study selects two typical cases:"China southern emergency incident" and "Nanjing little nurse was hit incident", the collection and organization methods are put forward to complete the social public opinion information collection, pretreatment, extraction and storage, which verifies the feasibility and effectiveness of the collection method based on the demand and the organization method based on ontology.