Abstract
Purpose
The increasing concern over mental health issues, particularly depression in children and adolescents, has prompted research into accurate methods of diagnosis. However, previous studies have been limited in terms of data and age groups, hindering their effectiveness. Therefore, this study aims to use machine learning algorithms with the latest data from Australia to improve accuracy in identifying depression in young people. It also highlights the importance of early detection and explores how depressive symptoms differ across different age groups. Additionally, the study provides insights into the likelihood of depression in individuals of various ages.
Methods
Machine learning algorithms were employed to identify the onset, persistence, and development of depression in children and adolescents. Three algorithms, random forest, support vector machine, and logistic regression, have been chosen for their ability to handle complex data, capture correlations, and make accurate predictions. These algorithms were applied to the recent longitudinal dataset from Australia, and their performance was compared.
Results
The study found that using the random forest algorithm produced noteworthy results in terms of accuracy, precision, recall, and F1 score for diagnosing depression. The random forest model achieved an impressive accuracy of 94% and weighted precision of 95%. Additionally, logistic regression was employed to measure the likelihood of depression, resulting in an accuracy rate of 89% and a weighted precision of 91%.
Conclusion
The study findings suggest that developing a predictive model that comprehends the characteristics and trends linked to mental illness during different stages of children and adolescents’ lives has substantial potential for promoting early intervention and identifying individuals who might be at risk of developing mental illness in the future.