Detection of fraud and economic risks in individual entrepreneurship in Kazakhstan using machine learning methods

A. S. Shainurov, A. B. Zhanys, Lazzat Kairatovna Dyusembayeva, Zh.O. Ryskulova

2026

Closed access Metadata (CSL-JSON)
DOI 10.58319/26170493_2026_2_144
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

In today's context of economic digitalization and expanding entrepreneurial activity, fraud detection and economic risk management in the individual entrepreneurship sector are particularly pressing. The growth in electronic payments, remote financial transactions, and digital services creates new opportunities for business development, but also increases the likelihood of financial irregularities, tax abuse, and fraud. This increases the need for modern analytical tools capable of ensuring the timely detection of suspicious transactions and minimizing economic losses. The aim of this study is to analyze the feasibility of applying machine learning methods to fraud detection and economic risk assessment in the activities of individual entrepreneurs in Kazakhstan. The methodological basis of the study is based on comparative and systemic analysis methods, as well as machine learning algorithms, including logistic regression, decision trees, random forests, gradient boosting, and neural networks. The paper examines the specifics of using big data and digital traces of entrepreneurial activity to develop risk prediction models and detect anomalous behavior in businesses. The study's results demonstrate that the use of machine learning technologies significantly improves the accuracy of fraud detection compared to traditional control methods based on manual verification and fixed rules. The use of intelligent algorithms facilitates the automation of financial transaction monitoring, reduces false positives, and improves risk management efficiency. However, issues related to the quality of source data, the protection of confidential information, and the need to improve the regulatory framework for digital security were identified. The practical significance of the study lies in the potential use of the developed approaches by government agencies, financial institutions, and businesses to increase business transparency, prevent financial violations, and strengthen the economic security of the Republic of Kazakhstan.