DEVELOPMENT OF ALGORITHM FOR ASSESSMENT OF RESULTS OF ACTIVITY OF POLICE DEPARTMENS WITH THE USE OF BINARY CHOICE MODELS
I. V. Goroshko, S. H. Gonov
This paper proposes a technological algorithm for evaluating the results on the basis of discrete choice models, which allows to evaluate the activities of the police departments. The mechanism of identification and evaluation of the parameters of logit and probit models, as well as methods for assessing the quality and significance are considered. The possibility of application of these models for the analysis of results of activity of the police departments is proved. In the article the technological algorithm of the analysis on the basis of models of discrete choice allowing to estimate operational activity of the police departments in the sphere of ensuring a law and order in public places is offered. The subject of the research is the methods and models of analysis of the results of the activities of the police departments of the Ministry of internal Affairs of Russia. The methods of multidimensional statistical analysis of data, as well as the mathematical apparatus of discrete choice models were used in the study. The mechanism of identification and estimation of parameters of logit and probit models on the basis of the maximum likelihood estimation is considered. Proposed methods for evaluating the significance of the parameters of the regression equations, and methods of evaluating model quality. The analysis of statistical dataset using logit – and probit- models has led to the conclusion about the greatest impact of performance indicators with higher coefficients in the system of performance evaluation. Modeling showed the possibility of using these models in the search for quantitative laws of the studied process. For the first time, using mathematical models of discrete choice, an approach to the analysis and evaluation of crime in public places was developed, on the basis of which new solutions were proposed to evaluate the results of operational and service activities of the police departments, which allows improving the quality of information and analytical work.
Keywords: algorithm, model of binary choice, evaluation, maximum likelihood estimation, logit-model.