|
KMID : 0897520210260020086
|
|
Journal of Korean Association of Social Psychiatry
2021 Volume.26 No. 2 p.86 ~ p.93
|
|
|
Preliminary Data of Machine-Learning Approach in Prediction of Depression in Juvenile Detention Center
|
|
Son Dong-Hun
Chang Jhin-Goo
Song Hoo-Rim
Lee Su-Young
Lee Seung-Hoon
Kim Hyun-Soo
Hong Min-Ha
|
|
|
Abstract
|
|
|
Objectives : It is well known that mental health problems and crime are highly related to youth crime, but there is little research on the mental health of young offenders in Korea. Furthermore, research on the application of machine learning in the mental health of children and adolescents is still novel. This preliminary study aims to investigate whether it is appropriate to apply machine learning algorithms to predict depression among female adolescent inmates.
Methods : The subjects were 87 young females in Cheongju Juvenile Center. A questionnaire was distributed to the subjects to gather their demographic information and crime-related information, as well as their adverse childhood experiences and Beck depression inventory scores using self-reported scale questionnaires. Based on the collected information, six models (logistic regression, random forest, supportive vector machine, decision tree, nearest neighbor, Adaboost) that can predict depression were created to compare the predictive performance between models using machine learning techniques.
Results : Results showed that 29 victims (25.7%) met the criteria of PTSD and 19 victims (16.8%) met the rigid criteria of PTSD. But, according to the subscales, 41 victims (36.3%) were diagnosed as PTSD. Victims with PTSD had more serious depression, anxiety, sleep disturbance, anger, social withdrawal and life stresses.
Conclusion : This study identified the current mental health status of female inmates with high accuracy by applying machine learning techniques to predict depression. The applicability of machine learning techniques to the management and surveillance of mental health in vulnerable groups was also highlighted.
|
|
|
KEYWORD
|
|
|
Adolescent, Depression, Machine learning
|
|
|
FullTexts / Linksout information
|
|
|
|
|
|
Listed journal information
|
|
|
|