AAI_2025_Capstone_Chronicles_Combined

MENTAL HEALTH RISK DETECTION USING ML​

​

​

​

​

8

Note. Chi-square scores represent how strongly each feature is related to the target variable. Larger values imply a greater dependency, indicating that the feature is likely to play a more significant role in predicting the correct classification.

Background and Model Selection Logistic Regression remains a foundational tool in epidemiology and clinical modeling due to its simplicity, transparency, and ease of interpretation. In mental health research during COVID-19 lockdowns, logistic regression achieved area under the curve (AUC) scores of 0.73–0.76 when modeling psychological instability based on isolation and stress (Alkhamees & A.M., 2021). Its interpretability makes it an ideal baseline model for our risk classification system.

208