Using Boosted Machine Learning to Predict Suicidal Ideation by Socioeconomic Status among Adolescents

Area of Science:

• Adolescent mental health research
• Machine learning applications in psychology
• Socioeconomic determinants of health

Background:

• Adolescent suicidal ideation is a significant public health concern.
• Understanding the interplay between socioeconomic status (SES) and mental health is crucial.
• Machine learning offers novel approaches to identify complex risk factors.

Purpose of the Study:

• To identify key risk factors for suicidal ideation in adolescents.
• To investigate the relationship between these risk factors and varying levels of socioeconomic status (SES).
• To leverage machine learning for robust risk factor identification and validation.

Main Methods:

• Analysis of a large dataset comprising 54,948 adolescent participants.
• Risk factor identification stratified by suicidal ideation status and three socioeconomic status (SES) levels.
• Validation of risk factor influence using synthetic minority over-sampling technique (SMOTE) and XGBoost algorithms.

Main Results:

• Adolescents experiencing suicidal thoughts reported higher levels of sadness, stress, and anxiety, and lower happiness.
• In high SES adolescents, academic achievement emerged as a primary risk factor for suicidal ideation.
• For low SES adolescents, emotional factors like stress and anxiety were the most significant contributors to suicidal ideation.

Conclusions:

• Socioeconomic status (SES) is a critical determinant of adolescent mental well-being.
• Interventions aimed at improving SES in adolescence could alleviate negative emotions.
• Enhancing SES may serve as a strategy to mitigate the risk of suicidal ideation in young people.