Research On Key Technology Of Knowledge Base Construction In Public Security Domain

With the acceleration of the process of informatization,the intelligence-led policing has gradually become the main mode of public security work.While the police are analyzing the information,they need to integrate their own business data,data from other industry and Internet together to find clues and information timely so that they can make accurate judgment on current situation and take corresponding measures in a timely manner.But there are still some questions during their analyzation,such as the poor efficiency caused by the traditional methods based on human model.This is due to the lack of a standardized knowledge base for intelligence and information processing in the domain of public security.After studying the results of the current domestic knowledge base,in view of the current status of public security work,this paper puts forward a method to build the knowledge base in the field of public security starting from the needs of public security work,and the key technologies are researched,The main work includes the following aspects:Firstly,a construction model of knowledge base on public security is proposed.In view of the requirements of information standardization and information analysis in public security work,two modules of knowledge acquisition and knowledge discovery are highlighted in the knowledge base model.Secondly,a drug type ontology is established.Taking the drug-related cases as an example,the Protégé ontology construction tool is used to construct the drug type ontology,and the information identification in the drug-related cases is standardized to facilitate the sharing of information and lay the foundation for the information extraction.Thirdly,complete the work of information extraction.In the framework of GATE information extraction platform,combined with the characteristics of drug-related cases and the key concerns of public security work,the platform plug-in was added and modified,the JAPE rules are written,the extraction of information on the collection of drug-related corpus is completed.The effectiveness of the work was evaluated using the accuracy rate,recall rate and F value.Finally,knowledge discovery based on decision tree algorithm and neural network algorithm is proposed.Through the training of information extraction results,two classifiers are constructed,and the knowledge reasoning of certainty and uncertainty is carried out respectively.The results of classification are compared using the test data and multiple indexes.At the end of this paper,the research process is summarized and the next step is further studied.