Kunstig intelligens kan forudsige selvmord

Research. Baseret på hospitalsindlæggelser har forskere på Vanderbilt University Medical Center bevist, at man kan forudse sandsynligheden for at en person vil begå selvmord.

Ved hjælp af maskinlærings-algoritmer har Colin Walsh, en dataanalytiker på Vanderbilt University Medical Center bevist, at man kan forudsige selvmordsforsøg med 80-90% sandsynlighed for, at en person vil begå selvmord indenfor for to år og 92% sandsynlighed, når det gælder forudsigelser inden for den næste uge.

Forudsigelserne er baseret data fra amerikanske hospitaler – inclusive alder, køn, postnummer, medicinforbrug og tidligere diagnoser. Walsh og hans hold samlede data fra 5167 patienter fra Vanderbilt University Medical Center som blev indlagt med tegn på selvskade eller selvmords-tilbøjelighed. that had been admitted with signs of self-harm or suicidal ideation.

Læs hele historien i Quartz

Abstract on the research: Traditional approaches to the prediction of suicide attempts have limited the accuracy and scale of risk detection for these dangerous behaviors. We sought to overcome these limitations by applying machine learning to electronic health records within a large medical database. Participants were 5,167 adult patients with a claim code for self-injury (i.e., ICD-9, E95x); expert review of records determined that 3,250 patients made a suicide attempt (i.e., cases), and 1,917 patients engaged in self-injury that was nonsuicidal, accidental, or nonverifiable (i.e., controls). We developed machine learning algorithms that accurately predicted future suicide attempts (AUC = 0.84, precision = 0.79, recall = 0.95, Brier score = 0.14). Moreover, accuracy improved from 720 days to 7 days before the suicide attempt, and predictor importance shifted across time. These findings represent a step toward accurate and scalable risk detection and provide insight into how suicide attempt risk shifts over time.

Get the full research

Comments are closed.

Password Reset
Please enter your e-mail address. You will receive a new password via e-mail.