Using GraphSage to predict admissions of suicide attempts

Hospitals require predictive tools to anticipate mental health crises, such as suicide attempts,

and to optimise their response to these emergencies. Despite prior research identifying potential influencing

factors, developing accurate predictive models remains a significant challenge. Leveraging advancements

in artificial intelligence (AI), this study introduces a predictive model based on a Graph Neural Network

(GNN), specifically the GraphSAGE algorithm. The model integrates data on suicide attempt admissions

and environmental variables from a hospital in Catalonia to capture complex relationships and patterns.

 

The Results: highlight the potential use of GNN solutions in this particular use case to provide actionable insights,

enabling hospitals to improve resource management and enhance their response to mental health crises.