QUANTUM SUPPORT VECTOR MACHINES: DEVELOPING VARIATIONAL ALGORITHMS FOR DATA CLASSIFICATION IN TECHNICAL, NATURAL, AND SOCIO-ECONOMIC SYSTEMS

Abstract

The paper presents research results on the application of the quantum support vector machine method in various domains, including technical, natural, and socio-economic systems. To explore the utilization of the quantum approach in technical information systems, the problem of malicious traffic classification was investigated. For natural systems, the classification of water quality was addressed, while customer churn classification was tackled in economic systems.

In the case of malicious traffic classification, the quantum model demonstrated superior effectiveness compared to classical methods, albeit with slightly longer training times due to limited availability of quantum computers. Considering the potential of quantum computing in processing large volumes of data and complex analytical tasks, its application in cybersecurity can contribute to the detection and prediction of cyber-attacks, network and system protection against threats, as well as the identification of vulnerabilities and the establishment of mitigation mechanisms.

Regarding ecological research, classical methods and the quantum classifier showcased similar classification quality, but the results did not reach the benchmark level. To enhance the models, one can leverage improved classification methods or combine classical and quantum approaches to achieve more precise results. In the context of customer churn prediction in financial tasks, the quantum support vector machine method outperformed classical methods, highlighting its significant potential for yielding high-quality results in the future. To further develop such applications, algorithm refinement, increased availability of quantum resources, and the active utilization of new approaches for training and model optimization are necessary.