| With the depth development of Police Performance Appraisal, a large number ofdata is accumulated. It is feasible to conduct early warning analysis on theperformance of public security through analysis and processing of such a largenumber of valuable historical experience data. Moreover, Police Performance EarlyWarning can accelerate feedback speed and improve the efficiency of the performanceappraisal.In this paper, first of all, on the basis of in-depth study of the currentperformance appraisal method used by the public security departments, the staticmodel and dynamic model of the Public Performance Appraisal are constructed. Thestatic model is mainly about all objects and attributes in Public Performance Appraisaland the dynamic model research process and state control of Public PerformanceAppraisal. Then, research and analyze the model of Public Performance EarlyWarning, including single indicator warning and overall performance warning, at thesame time, design a feasible warning realization mechanism. The system of PublicPerformance Appraisal with Public Performance Early Warning function, which isfull-featured and standardized of implementation process, is realized. And then focuson the implementation of the method of Public Performance Early Warning based onthe historical data. This paper aims at proposing an improved decision treeclassification algorithm named SNID3based on ID3algorithm and Rough Set Theoryto raise the accuracy of early warning. The datasets used by experiments come fromthe Police Performance Appraisal System and the UCI Machine Learning Repositoryrespectively. The comparative analysis of the experimental results shows that theSNID3algorithm is more accurate and stable.At present, the system of Public Performance Appraisal researched anddeveloped by the author and others has been put into use. The platform is operatingnormally which means the police performance evaluation and early warning methodsboth achieve the desired goal. |
PDF Full Text Request