Open AccessPractical aspects of statistical modeling and forecasting of crime based on time series data
Author(s) -
A. M. Terekhov,
Sergey Kuvychkov,
N. T. Mironov,
Sergey Smirnov
Publication year - 2021
Publication title -
ûridičeskaâ nauka i praktika
Language(s) - English
Resource type - Journals
ISSN - 2078-5356
DOI - 10.36511/2078-5356-2021-3-91-97
Subject(s) - econometrics , time series , multivariate statistics , regression analysis , statistical model , linear regression , series (stratigraphy) , computer science , regression , statistics , mathematics , paleontology , biology
The aim of the work is to implement regression models of crime in practice, which will show the complex impact of various socio-economic factors on the state of crime in Russia, and determine the forecast for the near future. The result of this study is the construction of a linear multivariate regression model and a predictive ARMA model with three forecast options. Based on the correlation analysis, the factors that have the strongest relationship with the variable that characterizes crime are determined. The multiple regression model showed the interrelated influence of individual factors on the state of crime in the framework of the developed equation. The quality of the obtained models is confirmed by evaluation tests for statistical significance. The initial data of the study were: official statistical data of the Russian statistics for 1992–2020, an array of data from more than 500 observations was used.