employee
East-Siberian Institute of the Ministry of Internal Affairs of Russia (Department of Information Technology, Associate Professor)
employee
Irkutsk, Russian Federation
employee
Irkutsk, Russian Federation
Introduction: It is noted that criminological forecasting based on a trend model, having a number of indisputable advantages in comparison with other types of criminological forecasting, has significant drawbacks – the inability to ensure the specified accuracy of the forecast, as well as limited forecasting capabilities in case of insufficient accuracy of the description of the observed data. The article explores the possibilities of improving the quality of data description in the construction of trend models. Materials and methods: the study is based on reference data on the state of crime and methods of mathematical statistics. The Results of study: an approach to criminological forecasting based on a trend with insufficient quality of the description of the initial data is proposed, based on a change in the functional dependence of the model and a decrease in the influence of abnormal data. The proposed approach is implemented to solve the problem of short-term forecasting of the number of registered crimes in the Republic of Buryatia. It was possible to improve the quality of the data description by 25% and obtain a fairly accurate interval estimate of the forecast. It is also shown that interval forecasting provides more complete and reliable information for decision-making and risk assessment, therefore it is preferable to point estimation of the forecast. Findings and Conclusions: based on the conducted research, the authors consider it possible to recommend to improve the quality of the description of the data of the trend model: analysis of the quality of the description of data in various forms of functional dependence of the trend model by comparing the coefficients of determination and evaluating the accuracy of the forecast; verification of the presence of abnormal data with their subsequent exclusion from consideration or reduction of their impact by changing the observation interval.
criminological forecasting, coefficient of determination, abnormal data
1. Ayvazyan, S.A., Mkhitaryan, V.S. Applied statistics. Fundamentals of Econometrics: Textbook for universities: In 2 vols. 2nd ed., ispr. Vol.1: Ayvazyan S.A., Mkhitaryan V.S. Probability theory and applied statistics. M.: UNITY-DANA, 2001. 656 p.
2. Demenchenok, O. G., Baranov, S. A. Choosing a model for criminological forecasting based on trend // Bulletin of the East Siberian Institute of the Ministry of Internal Affairs of Russia : scientific and practical journal. Irkutsk : East Siberian Institute of the Ministry of Internal Affairs of Russia. 2023. No. 2 (105). pp. 131-141.
3. Gorbunov, V. K. Mathematical modeling of market demand : a textbook / V. K. Gorbunov. - 2nd ed., reprint. and add. - Saint Petersburg : Lan, 2022. - ISBN 978-5-8114-3256-1. - Text : electronic // Lan : electronic library system. - URL: https://e.lanbook.com/book/213152 (accessed: 11.07.2023). - Access mode: for authorization. users.
4. Afanasyev, V. N. Time series analysis and forecasting : textbook / V. N. Afanasyev. - Saratov : AI Pi Ar Media, 2020. - 310 p. - ISBN 978-5-4497-0269-2. - Text: electronic // Digital educational resource IPR SMART: [website]. - URL: https://www.iprbookshop.ru/90196.html (accessed: 11.07.2023). - Access mode: for authorization. users.
