| DOI LINK: https://doi.org/10.59671/sOzId |
| Paper ID:sOzId |
| Volume:50 |
| Issue:3 |
| Title:Predictive Analytics in Pediatric Mental Health: A Machine Learning Approachto Depression Detection by Physical Activity and Behavioral Health Indicators |
| Abstract:Background: Depression in children and adolescents poses significant public health challenges, impacting individual and societal well-being. Traditional methods for diagnosing and predicting depression often rely on symptomatic assessments post-onset, delaying intervention. Recent advancements in machine learning (ML) provide new opportunities for early prediction and prevention. Methods: This study leveraged data from the U.S. National Health Interview Survey(NHIS) from 2004 to 2014, involving 27,642 participants aged 4-17. We employed machine learning models, including XGBoost, Random Forest, Decision Tree, and Bagging Classifier, to analyze a vast array of physical activity and health behavior data. These models were evaluated for their accuracy in predicting mental health indicators. Results: The XGBoost model demonstrated notable accuracy (77.5%) in forecasting mental health scores, closely followed by the Random Forest classifier (77.4%). The models were further assessed for their ability to classify varying degrees of mental anxiety, with the redefined categories of 'Mild', 'Moderate', and 'Severe'. The retrained XGBoost model showed enhanced accuracy across these categories (81.7%- 85.7%), with AUC values indicating reliable differentiation between mental health states. Discussion: This study expands the scope of ML in predicting depression, highlighting the intricate relationship between diverse health behaviors and mental health. The predictive accuracy of the models underscores the potential of M Linearly detection and intervention for depression. Future research should focus on refining these models for broader application and exploring their utility in real-world clinical settings. |
| Keywords:Child and Adolescent Depression, Machine Learning in Healthcare, Health Behavior Data Analysis, Predictive Modeling in Mental Health. |
| Authors:Tianyu Gao, Deyi Liang, Xia(Sebastiane) Chen, Xiaoyu Tao |
| Paper PDF Link: View full PDF |