Machine learning can provide strong predictiv

picture: Fig 7. The high 10 most essential questions for males vs females.
view extra 
Credit: Weller et al., 2021, PLOS ONE, CC-BY 4.0 (

Researchers have developed a brand new, machine learning-based algorithm that reveals excessive accuracy in figuring out adolescents who’re experiencing suicidal ideas and habits. Orion Weller of Johns Hopkins University in Baltimore, Maryland, and colleagues current these findings within the open-access journal PLOS ONE on November third, 2021.

Decades of analysis have recognized particular danger elements related to suicidal ideas and habits amongst adolescents, serving to to tell suicide prevention efforts. However, few research have explored these danger elements together with one another, particularly in giant teams of adolescents. Now, the sphere of machine learning has opened up new alternatives for such analysis, which might in the end enhance prevention efforts.

To discover that chance, Weller and colleagues utilized machine-learning evaluation to knowledge from a survey of highschool college students in Utah that’s routinely carried out to observe points similar to drug abuse and psychological well being. The knowledge included responses to greater than 300 questions every for greater than 179,000 highschool college students who took the survey between 2011 to 2017, in addition to demographic knowledge from the U.S. census.

The researchers discovered that they might use the survey knowledge to foretell with 91 p.c accuracy which particular person adolescents’ solutions indicated suicidal ideas or habits. In doing so, they had been capable of establish which survey questions had probably the most predictive energy; these included questions on digital media harassment or threats, at-school bullying, severe arguments at residence, gender, alcohol use, emotions of security in school, age, and attitudes about marijuana.

The new algorithm’s accuracy is greater than that of beforehand developed predictive approaches, suggesting that machine-learning might certainly enhance understanding of adolescent suicidal ideas and habits—and will thereby assist inform and refine preventive packages and insurance policies.

Future analysis might develop the brand new findings through the use of knowledge from different states, in addition to knowledge on precise suicide charges.

The authors add: “Our paper examines machine learning approaches utilized to a big dataset of adolescent questionnaires, as a way to predict suicidal ideas and behaviors from their solutions. We discover strong predictive accuracy in figuring out these in danger and analyze our mannequin with current advances in ML interpretability. We discovered that elements that strongly affect the mannequin embrace bullying and harassment, as anticipated, but additionally elements of their household life, similar to being in a household with yelling and/or severe arguments. We hope that this examine can provide perception to tell early prevention efforts.”

Method of Research
Computational simulation/modeling

Subject of Research

Article Title
Predicting suicidal ideas and habits amongst adolescents utilizing the chance and protecting issue framework: A big-scale machine learning method

Article Publication Date

COI Statement
The authors have declared that no competing pursuits exist.

Disclaimer: AAAS and EurekAlert! should not chargeable for the accuracy of stories releases posted to EurekAlert! by contributing establishments or for using any data by means of the EurekAlert system.

Recommended For You