Regression Models for the AI Gaming Chatbot for Learning Programming Based on Wordle-Type Puzzles

Automation and Remote Control. 2025. Vol. 86. No. 5. P. 417–431.

A. N. Varnavsky

Programming is one of the most important skills of the 21st century. However, for many students, learning programming is a rather complex process. In such cases, it is important to maintain the interest and involvement of students in the learning process. It is believed that digital games can solve this problem. One of the types of games that are well suited for the field of computer science are puzzles, which are aimed, among other things, at developing cognitive abilities. The purpose of the work is to develop models, an algorithm for the operation and structure of a game chatbot with artificial intelligence for learning programming using Wordle-like puzzles. Game Wordle was chosen due to its worldwide popularity and adapted as a gaming chatbot for use in learning programming. Artificial intelligence in the chatbot is necessary to control the appropriateness and appropriate time of its use, as well as adaptively forming the difficulty level of puzzle. Based on the data collected as a result of using a non-intelligent gaming chatbot, regression models of the influence of student indicators on the level of interest and difficulty of puzzles offered by the gaming chatbot were built. The developed models formed the basis of the algorithm of operation and structure of the gaming chatbot with artificial intelligence. When using an intelligent gaming chatbot, it is possible to further train the models and adjust the previously obtained coefficient values.

Keywords: обучение программированию, Game Artificial Intelligence, игровой искусственный интеллект, gaming chatbot, игровой чат-бот, learning to program

Publication based on the results of: Разработка и исследование эффективности использования виртуального игрового педагогического агента для обучения программированию и развития навыков hard и soft skills (2022)