Mining relationships between mental health, academic performance and human behaviour

摘要

The relationships among human behavior, mental health and academic performance have attracted researchers in psychology and education for many years. With the increase of the embedded sensor, the mobile phone can collect the data of the users' behavior from different aspects and different dimensions continuously and unconsciously. Focusing on behavioral relationship analysis, this paper presents a new method based on genetic algorithm to efficiently discover behavioral correlates of mental health (i.e., depression, stress, loneliness, self-perceived success) and academic performance. We conduct the experiments based on the open dataset called “StudentLife”. Ultimately, we discover interesting relationships which have never been considered before. We believe this approach provides a novel and effective way to mine massive relationships from mobile contextual data in the future.