Department of Psychology, Macquarie University, Sydney, NSW, Australia.
Department of Computing, Macquarie University, Sydney, NSW, Australia.
Suicide Life Threat Behav. 2021 Jun;51(3):455-466. doi: 10.1111/sltb.12719. Epub 2020 Nov 13.
Identifying predictors contributing to suicide risk could help prevent suicides via targeted interventions. However, using only known risk factors may not yield accurate enough results. Furthermore, risk models typically rely on suicidal ideation, even though people often withhold this information.
This study examined the contribution of various predictors to the accuracy of six machine learning models for identifying suicidal behavior in a prison population (n = 353), including borderline personality disorder (BPD) and antisocial personality disorder (APD) criteria, and compared how excluding data about suicidal ideation affects accuracy.
Results revealed that gradient tree boosting accurately identified individuals with suicidal behavior, even without relying on questions about suicidal ideation (AUC = 0.875, F1 = 0.846). Furthermore, the model maintained this accuracy with only 29 predictors. Meeting five or more diagnostic criteria of BPD was an important risk factor for suicidal behavior. APD criteria, in the presence of other predictors, did not substantially improve accuracy. Additionally, it may be possible to implement a decision tree model to assess individuals at risk of suicide, without focusing upon suicidal ideation.
These findings highlight that modern classification algorithms do not necessarily require information about suicidal ideation for modeling suicide and self-harm behavior.
确定导致自杀风险的预测因素有助于通过针对性干预预防自杀。然而，仅使用已知的风险因素可能无法得出足够准确的结果。此外，风险模型通常依赖于自杀意念，尽管人们经常隐瞒这方面的信息。
本研究探讨了各种预测因素对六种机器学习模型在监狱人群中识别自杀行为的准确性的贡献（n=353），包括边缘型人格障碍（BPD）和反社会人格障碍（APD）标准，并比较了排除自杀意念数据对准确性的影响。
结果表明，梯度提升树准确地识别出有自杀行为的个体，即使不依赖于自杀意念的问题（AUC=0.875，F1=0.846）。此外，该模型仅使用 29 个预测因素即可保持这种准确性。符合 BPD 的五个或更多诊断标准是自杀行为的一个重要危险因素。在存在其他预测因素的情况下，APD 标准并不能显著提高准确性。此外，有可能实施决策树模型来评估有自杀风险的个体，而不必关注自杀意念。
这些发现强调，现代分类算法不一定需要关于自杀意念的信息来对自杀和自残行为进行建模。
