Learning Success Prediction Model for Early Age Students Using Educational Games and Advanced Data Analysis (CROSBI ID 70381)
Prilog u knjizi | ostalo | međunarodna recenzija
Podaci o odgovornosti
Tolić, Antonio ; Mršić, Leo ; Jerković, Hrvoje
engleski
Learning Success Prediction Model for Early Age Students Using Educational Games and Advanced Data Analysis
The early years of a child's life greatly affects education potential and furthermore potential of educational achievements in adulthood. The brain de-velops faster during early age, and missed cognitive opportunities in that period are difficult to make up for. In this research, we have developed a machine learn-ing model based on the data points obtained from the educational game, with aim to predict how many attempts are necessary for an individual child to complete the task or assessment as part of educational game. In-game assessments are based on the skills that the child already possess and those developed while playing the game. Training of the machine learning model is based on collected and processed data points (features), while model interconnections are related to the factors of the child's cognitive growing up process. Model performance bench-marks are elaborated in results and conclusion section of the paper as quality measures of the forecast indicators.
artificial intelligence, machine learning, XGBoost, CatBoost, QWK, data preparation and processing, educational game, learning success prediction model, quadratic weighted kappa, extreme gradient boosting
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
1-12.
objavljeno
10.1007/978-3-030-68154-8_61
Podaci o knjizi
Proceedings of the 3rd International Conference on Intelligent Computing and Optimization 2020 (ICO 2020)
Pandian Vasant, Ivan Zelinka, Gerhard-Wilhelm Weber
Zürich: Springer
2021.
978-3-030-68154-8