Detecting suicidal risk using MMPI-2 based on machine learning algorithm.

Area of Science:

• Psychological assessment
• Machine learning in mental health
• Clinical psychology

Background:

• The Minnesota Multiphasic Personality Inventory-2 (MMPI-2) is a standard tool for psychological assessment.
• Early detection of psychological maladjustment and adaptation levels is crucial in various settings.
• Assessing suicidal risk is a critical component of mental health evaluation.

Purpose of the Study:

• To evaluate the utility of the MMPI-2 in assessing suicidal risk.
• To determine the accuracy of machine learning models in predicting suicidal ideation and attempts using MMPI-2 data.
• To analyze the predictive power of MMPI-2 scales for suicide risk in a large college student population.

Main Methods:

• Analysis of 7,824 datasets from college students.
• Utilized MMPI-2-Restructured Clinical Scales (MMPI-2-RF) and Mini International Neuropsychiatric Interview (MINI) suicidality module data.
• Applied Random Forest and K-Nearest Neighbors (KNN) machine learning techniques.

Main Results:

• Random Forest model achieved 92.9% accuracy for suicidal ideation and 95% for suicide attempts, with AUCs of 0.844 and 0.851.
• KNN model achieved 91.6% accuracy for suicidal ideation and 94.7% for suicide attempts, with AUCs of 0.722 and 0.639.
• Machine learning with MMPI-2 data demonstrated reliable accuracy in classifying and predicting suicidal ideation and past attempts.

Conclusions:

• Machine learning approaches using MMPI-2 data are effective for large-scale prediction of suicidal ideation and attempts.
• The MMPI-2, when analyzed with machine learning, provides a reliable method for identifying individuals at risk for suicide.
• This study validates the use of MMPI-2 in conjunction with advanced analytical techniques for robust suicide risk assessment.