Machine learning for suicide prediction and prevention: Advances, challenges, and future directions (IN: Youth Suicide Prevention and Intervention, edited by J.P. Ackerman & L.M. Horowitz)

This chapter describes the role of machine learning in youth suicide prevention. Following a brief history of suicide prediction, research is reviewed demonstrating that machine learning can enhance suicide prediction beyond traditional clinical and statistical approaches. Strategies for internal and external model evaluation, methods for integrating model results into clinical decision-making processes, and ethical issues raised by building and implementing suicide prediction models are discussed. Finally, future directions for this work are highlighted, including the need for collaborative science and the importance of both data- and theory-driven computational methods.