МАШИНАЛЫҚ ОҚЫТУ КӨМЕГІМЕН ҚАЗАҚСТАНДАҒЫ СОТ ШЕШІМДЕРІН БОЛЖАУҒА АРНАЛҒАН ДЕРЕКТЕРДІ ҚҰРУ

Abstract

In this paper, a dataset for modeling and predicting the outcomes of court cases in Kazakhstan was developed and generated using machine learning methods. The generated data, based on articles of the "Code of the Republic of Kazakhstan", covers various criminal and administrative offenses. The dataset includes 100,000 cases with various attributes, such as the age of the accused, degree of guilt, mitigating and aggravating circumstances. This created a basis for developing predictive models applicable to the analysis and prediction of legal outcomes. The article discusses in detail the methods of data generation and the results of applying various machine learning algorithms to predict the outcomes of court cases. The relevance of the study is due to the need to automate the processes of law enforcement practice and improve the accuracy of predicting the outcomes of court cases in Kazakhstan. Traditional analysis methods require significant time and resource costs, and are complicated by the confidentiality of real data. The use of generated data based on the legislation of the Republic of Kazakhstan allows us to explore patterns and develop models for forecasting, which increases the efficiency of law enforcement practice. The aim of the study is to create a set of generated data for the analysis of court cases, as well as to develop and train machine learning models to predict the outcomes of cases based on the criminal and administrative legislation of Kazakhstan.