|
Data clustering analysis to identify high-risk Thai individuals for domestic violence and mental health problems using machine learning |
|---|---|
| รหัสดีโอไอ | |
| Creator | Wutthiphong Khuandin |
| Title | Data clustering analysis to identify high-risk Thai individuals for domestic violence and mental health problems using machine learning |
| Contributor | Chanida Kaewphet |
| Publisher | Mahasarakham University |
| Publication Year | 2569 |
| Journal Title | Journal of Science and Technology Mahasarakham University |
| Journal Vol. | 45 |
| Journal No. | 3 |
| Page no. | 378-388 |
| Keyword | Mental health, domestic violence, data clustering |
| URL Website | https://li01.tci-thaijo.org/index.php/scimsujournal |
| Website title | Journal of Science and Technology Mahasarakham University |
| ISSN | 1686-9664 (Print), 2586-9795(Online) |
| Abstract | Domestic violence and mental health problems remain critical public health concerns in Thailand, with an increasing prevalence observed during economic crises and the COVID-19 pandemic. Despite their significant societal impact, previous research has lacked in-depth analytical approaches utilizing data science to identify high-risk populations. This study aims to classify high-risk Thai individuals vulnerable to mental health issues and domestic violence by applying unsupervised machine learning techniques, specifically K-means, Hierarchical Clustering (HC), and Gaussian Mixture Models (GMMs). Internal evaluation metrics, including the Silhouette Score, Calinski-Harabasz Index, and Davies-Bouldin Index, were used to assess clustering performance. The dataset comprised 1,162 records of domestic violence offenders obtained from the Digital Government Development Agency (DGA) in Thailand. The findings indicate that HC achieved the highest performance (Silhouette Score = 0.429, Calinski-Harabasz Index = 61.790, and Davies-Bouldin Index = 1.034), effectively differentiating risk groups. The high-risk group was predominantly characterized by middle-aged males with mental health issues, substance abuse, and economic stress. This study demonstrates the potential of machine learning for identifying vulnerable populations. It provides insights that can inform the development of targeted prevention strategies, early warning systems, and evidence-based policymaking to mitigate domestic violence and promote sustainable mental well-being. |
