MACHINE LEARNING MODEL FOR PREDICTION OF SMARTPHONE ADDICTION

Abstract

Purpose: The primary objective of the present study is to predict the levels of smart-phone addiction and also to find the correlation between different smart phone activities, and their relationship across male and female users.
Methodology: The survey was conducted using a well-designed questionnaire which enquires about the usage of smartphone of an individual. College undergraduates (N = 115) participated in the survey and completed the questionnaire as part of their class requirements. The data thus collected is trained to form a machine learning model based on clustering.
Results: The findings significantly shows that males tend to use smartphones more than females to access books and e-books. that female has the largest count for possession of phones for more than 12 hours, whereas, male have the largest count for possession of their phones for less than 6 hrs. The results show that most of the male have their phone‟s battery last for a day, whereas for females the count of “yes” and “no” are almost equal. The whole population is categorized in 3 clusters such as Highly addicted group, moderately addicted group, non addicted group.
Conclusions: This prediction model certainly be highly useful for understanding the phone usage level and eventually predicting certain possible threats prevalent amongst addictive smartphone users