Machine learning in building cadet maladaptation forecasts

Relevance. Machine learning methods allow for high-accuracy classification and prediction of various conditions and outcomes in humans. The choice of the most optimal method is critical for researchers. The objectiveis to evaluate the effectiveness of machine learning methods in predicting cadet maladaptation.

Methods. The investigators examined 1822 cadets aged 18 to 27 years studying at the Federal State Higher Military Educational Institution “The Military Educational and Scientific Centre of the Navy “The Naval Academy named alter Admiral of the Fleet of the Soviet Union N.G. Kuznetsov”. The subjects were divided into 2 groups: normal (n = 1507) and maladaptation (n = 315). The examination was carried out using the Multifactorial Score for Adaptability Evaluation and the KR-3-85 intellectual development test. Statistical processing was carried out using the Stat Soft Statistica 10.0 software package. The normality of indicators was verified using the Kolmogorov–Smirnov test. Comparative analysis of indicators with normal distribution was assessed by Student’s t-test. Spearman rank correlation analysis was performed to assess multicollinearity of data. Mathematical modeling was carried out using neural networks, discriminant analysis, and the Bayesian algorithm. The effectiveness of the models was assessed by sensitivity and specificity parameters.

Results and discussion. Neural networks and Bayesian algorithm are powerful classification tools, allowing to reliably classify cadets with socio-psychological maladjustment. At the same time, the Bayesian algorithm is characterized by high sensitivity, whereas neural networks show by high specificity. Loss of data is a well-known disadvantage of discriminant analysis modeling. This, discriminant analysis failed to classify cadets with social and psychological maladjustment.

Conclusion. The use of machine learning will increase the efficiency of medical and psychological support for cadets. Neural networks are the optimal method to predict maladaptation.

1. Borisov D.N., Koluzov A.V., Serezhkin I.A. Vozmozhnosti razvitiya iskusstvennogo intellekta i bol’shih dannyh v oblasti zdorov’ya voennosluzhashchih [Possibilities for the development of artificial intelligence and big data in the field of military health]. Sostoyanie i perspektivy razvitiya sovremennoj nauki po napravleniyu “ASU, informacionno-telekommunikacionnye sistemy“ [State and prospects for the development of modern science in the direction of “ACS, information and telecommunication systems”: Scientific. Conf. Proceedings]. Anapa. 2020; 3:177–183. (In Russ.)

2. Degtyarenko K.A. Iskusstvennyj intellekt v medicine. Obzor 21 mezhdunarodnoj konferencii po iskusstvennomu intellektu v medicine [Artificial intelligence in medicine. review of the 21st international conference on artificial intelligence in medicine]. Aziya, Amerika i Afrika: istoriya i sovremennost’ [Asia, America and Africa: history and modernity]. 2023; 2(3):27–42. (In Russ.)

3. Lysova M.E., Kuznecov M.E. Nejronnye seti v medicine. Avtomatizaciya pri pomoshchi iskusstvennogo intellekta [Neural networks in medicine. Automation using artificial intelligence]. Dostizheniya nauki i tekhnologij – DNiT-11-2023 [Achievements of science and technology – DNiT-11-2023 : Scientific. Conf. Proceedings]. Krasnoyarsk. 2023:581–586. (In Russ.)

4. Lipskij D.L., Gura M.S., Luchkin I.V., YUsupov V.V., YAtmanov A.N. Primenenie mashinnogo obucheniya v medikopsihologicheskom soprovozhdenii voennosluzhashchih inostrannyh gosudarstv [Application of machine learning in medical and psychological support of military personnel of foreign states]. Zhivaja psihologija [Living psychology]. 2023; 10(4):15–24. (In Russ.)

5. Mel’nikova E.V. Glubokoe mashinnoe obuchenie v optimizacii nauchno-issledovatel’skoj deyatel’nosti [Deep machine learning in optimization of research activities]. Nauchno-tekhnicheskaya informaciya. Seriya 1: Organizaciya i metodika informacionnoj raboty [Scientific and technical information. Series 1: Organization and methodology of information work]. 2023; (2):8–13. (In Russ.)

6. Yasnickij L.N. Intellektual’nye sistemy [Intelligent systems]. Moscow. 2016. 221 p. (In Russ.)