Research And Implementation Of Entity Recognition System For Police Imformation Analysis

The police case has certain similarity and predictability.Through the analysis of historical police information,it is possible to combine the cases to speed up the crime resolution and prevent crimes.In the past,the police recorded the police information in specific files by writing on paper,and finally filed different case.But now,the police can use the online police information management system to complete the recording and retrieval of police cases.Although the police information management system has greatly improved the efficiency of recording and retrieval,only the description of the police event can not effectively complete the analysis of the police information,but also requires the multi-dimensional analysis of the information in the police situation,such as the person who reported the crime,the location of the crime,and the organization involved.Based on the above background,this paper studies how to extract the effective entity information in police information from the perspective of police information analysis,and designs and implements an entity recognition system for police information analysis.This paper takes the analysis of the police information as the goal,and conducts an in-depth study on the effective entities in the police information that are beneficial to the case analysis,and proposes an entity recognition methods suitable for the police information.In this paper,two methods,conditional random field and deep neural network combined with conditional random field,were tested and compared.Both methods have achieved high accuracy,recall rate and F1 value,but from the evaluation results of each category and the whole,it was concluded that the method based on deep learning was indeed better than the traditional model.The main work of this paper is as follows:Firstly,the relevant technologies and algorithms used in the system are introduced,including word representation,probability graph model,neural network and other theoretical bases.Secondly,through the analysis of police information,the extracted police location is subdivided,and the task of building the police entity recognition model is completed.Thirdly,the functional requirements and non-functional requirements of the system are analyzed.According to the results of the requirements analysis,the hierarchical structure and functional modules of the systems are designed,and the database related design is completed.Finally,the realization of the main function modules of the system is given,and the operation interface of the system is shown.The police entity recognition model completed in this paper has been tested and deployed to the internal system of the project partner.In half a year of use,the model runs well with high recognition,which is appreciated by the proj ect partner and proves the effectiveness of the model.