Analysis using spectral clustering to predict Internet gaming behaviours

As computers are becoming more powerful and with the availability of sophisticated data visualization software, the capability to extract knowledge from data for the benefit of individuals and society is imperative. Contemporary research barely started to reap the power of data science techniques to process and analyze data, and communicate results. Actually, research studies from various disciplines have been relying heavily on traditional statistical methods such as correlations and regressions. In this study, we propose a novel approach using clustering analysis to tackle the problem of the impact of Internet gaming. The special interest in Internet gaming is intensifying as it is spreading widely among students and as it is believed to have detrimental effects on their academic performance and sleep habits. We aim to identify patterns related to Internet gamers by using spectral clustering to determine the structure among Internet gaming, academic achievement, and sleep habits. Results show three distinctive clusters and a strong associations between Internet gaming disorder on one hand, and decline in both sleep hours and academic performance on the other hand.