Detection of Suicide Attempters among Suicide Ideators Using Machine Learning

Area of Science:

• Public Health
• Computational Psychiatry
• Data Science

Background:

• Suicide remains a significant public health concern globally.
• Identifying individuals at high risk of suicide attempts is crucial for timely intervention.
• Existing methods for risk assessment have limitations in predictive accuracy.

Purpose of the Study:

• To develop and validate machine learning models for predicting suicide attempts among individuals with suicidal ideation.
• To leverage a large-scale national health survey dataset for model development.
• To assess the performance of the predictive models using robust evaluation metrics.

Main Methods:

• Utilized data from the Korea National Health & Nutrition Examination Survey (KNHANES) on 5,773 individuals with suicide ideation.
• Applied Synthetic Minority Over-sampling Technique (SMOTE) for data balancing.
• Trained random forest models using recursive feature elimination and 10-fold cross-validation on a training set (n=1,858) and tested on a separate set (n=796).

Main Results:

• The developed machine learning model demonstrated high predictive performance in the test set.
• Achieved an area under the receiver operating characteristic curve (AUC) of 0.947.
• Attained a prediction accuracy of 88.9% for identifying suicide attempters.

Conclusions:

• Machine learning offers a powerful tool for predicting suicide attempts in individuals experiencing suicidal ideation.
• Integrated analysis of diverse suicide risk factors using AI can enhance risk stratification.
• This approach holds promise for proactive suicide prevention strategies.