Kratos: An Intelligent Model for Assisting People with Problematic Smartphone Use Through Context Histories Analysis

Abstract

Global smartphone usage has surged, becoming indispensable in people’s daily lives. Despite benefits, concerns arise about prolonged hyper-connected experiences. The excessive use of smartphones coupled with demographic and mental health-related risk factors can lead to problematic smartphone use (PSU), characterized as compulsive smartphone use disrupting daily life, work, and relationships. This article summarizes an Academic Master’s Dissertation introducing Kratos, a computational model designed to identify PSU through context awareness, context histories, machine learning, and ontology. The main scientific contribution of the Kratos model is the automatic PSU identification and intervention proposals using machine learning and ontological inferences. In the assessment of the model, the Kratos Dataset Simulator (KDS) generated simulations for 49 individuals for 30 days. The machine learning and ontology experiments occurred based on the KDS simulated dataset. The creation of smartphone use behavior profiles allowed the use of Manhattan Distance to identify the behavior as normal or PSU. A silhouette analysis allowed the validation of the consistency of the clusters after the behavior identification process occurred. The results demonstrated the model’s ability to consistently distinguish the smartphone use behaviors, correctly separating the clusters of behaviors. Based on the Machine Learning and Ontology results, Kratos recommends literature-based interventions for PSU behaviors. Thus, this research improves PSU identification and assistance through the proposed model.