Tools
Tools
Shaik, Saleem Pasha (2025) Innovative Use of Behavioural Data and Explainable AI for Early Gambling Addiction Detection. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (949kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (1MB) | Preview |
Abstract
Gambling addiction has grown as a behavioural and financial issue which is driven by the growing accessibility of online betting platforms and the difficulty in identifying addictive patterns early. This study proposes an Explainable AI-based framework that uses behavioural data analytics and deep learning to detect early signs of gambling addiction among online players. The dataset uses demographic, betting and some intervention data to model player behaviour. The approach starts with K-Means clustering to segment players into casual, moderate-risk and high-risk categories which has evaluated using Silhouette Score, Davies–Bouldin Index and Calinski–Harabasz Index to secure strong cluster quality. High-risk player data is further analyzed using LSTM, GRU, BiLSTM and BiLSTM with Cross-Attention models for time-series forecasting of betting patterns which is bee optimized through early stopping and adaptive learning rate scheduling. Performance has measured with the help of MSE, RMSE, MAE and model response metrics like latency and throughput. Results shows that BiLSTM with Cross-Attention provides superior predictive accuracy and balanced computational performance. The use of LIME explainability enhances transparency by securing model interpretability for responsible gambling. This study contributes a novel, interpretable predictive framework for early detection and prevention of gambling addiction using behavioural data and deep learning.
Actions (login required)
 |
View Item |