Research And Application Of Constraction Technology Of Social Network Based On Information Extraction

In recent years, the use of massive data analysis to assist in criminal investigations and countermeasures to prevent the occurrence of crime has become a significant challenges which law enforcement and intelligence agencies face. Statistical analysis of crime data, crime data structure of mining, social network analysis methods been widely studied.The social network analysis in terms of network characteristics for the analysis of crime increasingly recognized. By building criminal networks, with nodes that criminal elements, to represent the connection between analysis and mining information on criminals, psychological behavior, law, find hidden clues, digging out the important nodes in the criminal networks and the potential structure can assist criminal investigations. With the growing number of unstructured information, information analysis and the difficulty increases exponentially, the structured data to construct social networks approach which is in the past has been not suitable for a dynamic and vast amounts of information. Extract information on crime from the large amount of unstructured or semi-structured data and build visualization criminal networks, control the structure of criminal networks is great significance for dealing with criminal cases.This paper will use the combination of Ontology and information extraction technology to construct criminal networks, which is a new unstructured information-oriented extraction of construction method of criminal networks. Using information extraction technology to unstructured or semi-structured text, extract the contents need to build criminal networks, and save the structured data, such as XML documents or relational databases. Available to the social network visualization tools for visual display and social network analysis to help law enforcement solve the case. However, there are also still a lot of difficulties in the unstructured information extraction technology.Currently, information extraction system on statistical learning method depends on the specific areas of training text, and accuracy is not high. Information extraction system on pattern matching method, due to the limitations of its pattern-matching technology, is only for a particular relationship between entity type and entity extraction.This paper introduces of ontology-based information extraction technology. Through analysis of the demand for criminal networks, build suspects ontology, the ontology with instances of ontology construction, the final completion of information extraction and generate consistent with the structure of social network visualization XML document. Combining with crime data provided by research projects, the construction of social networks is judged the feasibility of building criminal networks, according to information extraction rate of recall and precision. Experimental results show that the proposed social network based on building information extraction technology that can effectively suspects from unstructured text, information, relationships and build the visual network, the detection of cases, help combat crime. On the basis of this framework can be extended body, the area of unstructured information more crime, the extraction of building area network.