Research And Implementation On Intelligence Analysis System Based On Content Analysis

With the development of information technology, Web2.0and mobile internettechnology has been growing rapidly, the world has entered a data-centric paradigm,that is the “Big data” era. Intelligence analysis in the background of big data is to findvaluable information for user in massive, complex and rapidly growing data. Thetraditional methods of intelligence analysis mainly depends on manual analysis andretrieval, and has the following problems:(1) the manual analysis cannot complete themassive intelligence material analysis tasks;(2) the retrieval analysis can only analyzesome keywords and ignore most of the content of the intelligence material;(3) theanalysis results are usually described in tables and text, the intelligence inherentknowledge and regularity cannot be showed obviously.This paper firstly reviews the status of related technology and introduces contentanalysis and its key technologies. Secondly, in order to solve the problem thattraditional intelligence analysis being unable to analyze the total content, proposes anintelligence analysis model based on content analysis; in order to solve the result-instability problem of the traditional K-Means algorithm, proposes an optimizedK-Means algorithm for intelligence analysis and the availability of the algorithm isverified; in order to solve the problem that tables and text description being unable toshow knowledge of intelligence obviously, propose a knowledge mapping constructingmethod based on stratification theory. Finally, uses the research results to a PoliceIntelligence Analysis SystemThe main contents of this paper are summarized as follows:(1) According to the requirement of big data, combined with content analysistheory, proposes and designs a typical three-tier structure of intelligence analysis model.The model takes content analysis As the basic framework, uses NoSQL, and conducts acomprehensive analysis of the content of the intelligence material.(2) Proposes an optimized K-Means algorithm for text clustering analysis,experimental verification shows that the algorithm can be used for intelligenceclustering and the purity has improved.(3) According to the information which mined out by the use of content analysis, proposes organization and visualization methods based on stratification mappingknowledge. The methods can replace the traditional charts and text and to show thephenomenon and the knowledge of the intelligence material in the form of mapping.(4)On the basis of the research above, applies for a Police Intelligence AnalysisSystem, designs and implements intelligence text preprocessing module, similaritycomputing on cases and automatic clustering module, builts the intelligence knowledgemap, discovers some knowledge of the intelligence material behind.