Machine Learning Solutions for Reducing Childhood Internet Dependency
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Problematic Internet Use is a growing public health concern, particularly among adolescents, with substantial implications for mental and physical health. This study intends to estimate the level of internet addiction using the Child Mind Institute (CMI) dataset, which comprises a complete variety of demographic, clinical, physical, and time-series actigraphy data. XGBoost, LightGBM, CatBoost, and TabNet, based on the Cohen’s Kappa as the main evaluation objective. Apart from 5-fold cross-validation for training and validation of our model, our method includes data preparation, feature engineering, model training, and validation. The findings reveal that XGBoost gives the maximum performance, with a training Kappa value of 0.80 and constant validation scores across folds, while TabNet al-so demonstrates strong performance with minimal overfitting. Analysis of feature importance suggests that psychological factors such as anxiety and impulsivity are the most significant predictors of PIU. These findings emphasize the potential of machine learning in identifying persons at risk and permitting targeted interventions. Future study will try to test the model on larger datasets and examine its application in real-world therapeutic and educational contexts to handle the growing worry of internet addiction among children.
Fields of Study & Subjects
Impact of Technology on Adolescents
Artificial Intelligence in Healthcare
Child Development and Digital Technology
Publication details
- Journal / source
- CRC Press eBooks
- Pages
- 541–548
- Publication date
- 31 March 2026
- Language
- EN
- Document type
- Book chapter