An Influence Factors Analysis and Prediction of Mental Disorder

Abstract

The prevalence of mental disorders continues to rise, necessitating increasingly stringent diagnostic requirements. This study conducted a comprehensive visualization analysis of raw data, performing correlation analyses between dependent variables and all independent variables, as well as pairwise relationships among all independent variables. The primary objective was to compare the performance of predictive models, such as logistic regression and random forest, in mental health diagnostics and to identify the most significant influencing factors for different disorders. The results indicate that factors associated with depression, bipolar I disorder, and bipolar II disorder are all correlated with emotional states. Aggressiveness emerged as the most prominent factor in bipolar I disorder, while suicidal ideation was identified as a significant factor in bipolar II disorder. Depression demonstrated strong correlations with self-loathing and self-harm tendencies. Symptomatology reflects alterations in emotional states. Among the models evaluated, the random forest algorithm demonstrated superior accuracy, stability, and sensitivity. Consequently, it demonstrates superior performance.