|
Abstract
|
OBJECTIVE: This study aimed to investigate the similarities and differences in risk factors for suicide among adult and adolescent women in South Korea and identify subtypes of suicidal ideation or suicide attempt in each group.
METHODS: Multifaceted data were collected and analyzed by linking survey and social media data. Interpretable machine learning models were constructed to predict suicide risk and major risk factors were extracted by investigating their feature importance. Additionally, subtypes of suicidal adult and adolescent women were identified and explained using risk factors.
RESULTS: The risk factors for adult women were primarily related to mental disorders, while those for adolescent women were primarily related to interpersonal experiences and needs. Two subtypes of suicidal adult women were one with high psychiatric symptoms and mental disorders of them and/or their families and the other with excessive social media use and high online victimization. Two subtypes of suicidal adolescent women were one with high psychiatric symptoms, high ACEs, and high social connectedness, and the other with frequent social media use, high online sexual victimization, and high social assurance.
CONCLUSIONS: These findings enable a stratified and targeted understanding of suicide in women and help develop customized suicide prevention plans in South Korea.
Language: en |