Гейміфікація процесу підтримки прийняття рішення щодо вибору професії підлітками

The object of this study is the process of forming recommendations for adolescents regarding their choice of profession based on performance in a multi-level, profession-oriented computer game. The problem addressed is the gamification of adolescents’ professional self-identification A multilayer perceptron with two sequential hidden layers of 32 and 16 nodes has been proposed to generate recommendations in real time. The task is formulated as a categorization problem using game performance data, including interest level, learning readiness, time spent, and the number of attempts for level completion. The neural network design allows integration with various games and expansion of the training dataset. Combining "time" and "number of attempts" ensures the accuracy and efficiency of the training process, mitigating evaluation bias caused by users guessing strategies through frequent attempts. Experts validate the game’s alignment with professional qualification requirements and assess adolescents’ readiness to acquire necessary competencies. Abilities are evaluated by comparing the gameplay parameters of users with those of specialists. The specialist’s results serve as benchmarks to normalize data, forming the basis for recommendations. Unlike previous approaches, this study accounts for both the time taken to complete each level and the number of attempts, ensuring fairer evaluations. The practical significance of the work is to offer users accessible tools for career decision-making through engaging in profession-oriented games. The neural network model has the potential for implementation in career guidance systems, provided the training dataset is expanded and the system undergoes additional testing under a trial mode for refinement and adaptation