Scholarlink

<p>Student mental health varies across regions and remains underexplored in low- and middle-income countries (LMICs). This study examines regional differences in mental health among university students at Shahjalal University of Science and Technology (SUST), Bangladesh. Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) identified three key factors from the General Health Questionnaire (GHQ-12): (i) negative emotions (insecurity = 0.95, worthlessness = 0.67, unhappiness = 0.54), explaining 19% variance, (ii) positive traits (resilience = 0.53, happiness = 0.69, enjoyment = 0.53), contributing 15%, and (iii) stress-related difficulties (insomnia = 0.65, stress = 0.61, academic difficulties = 0.45), accounting for 13%. The model demonstrated strong reliability (CFI = 0.975, TLI = 0.960, RMSEA = 0.066). Machine learning models detected regional differences in mental health. The stacking model (SM) achieved the highest predictive performance (accuracy = 98.0%, sensitivity = 88.0%, specificity = 100%). Other models performed moderately, including ensemble standardize (ESS) (accuracy = 74.3%) and voting classifier (VC) (accuracy = 70.7%). Dhaka students displayed higher concentration (1.57), resilience (1.75), and happiness (1.81), while Chattogram students reported greater enjoyment (1.83). Rajshahi (stress = 2.00) and Rangpur (insecurity = 3.00) students exhibited higher risks. Model accuracy peaked in Dhaka (98.0%) and dropped in Rangpur (74.2%). The stacking model prioritized accuracy over interpretability, highlighting the need for statistical models to enhance explanatory power. Findings emphasize targeted mental health interventions for students in rural areas to improve well-being in LMICs.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Regional mental health disparities among university students in Bangladesh: a comprehensive factor analysis and predictive modeling approach

  • 期刊论文

  • 发表日期：

    2025年4月29日

查看全文
  • Tahsin Tamanna,
  • Emon Barua,
  • Md. Najmul Kabir,
  • Zia Ahmed

摘要

Student mental health varies across regions and remains underexplored in low- and middle-income countries (LMICs). This study examines regional differences in mental health among university students at Shahjalal University of Science and Technology (SUST), Bangladesh. Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) identified three key factors from the General Health Questionnaire (GHQ-12): (i) negative emotions (insecurity = 0.95, worthlessness = 0.67, unhappiness = 0.54), explaining 19% variance, (ii) positive traits (resilience = 0.53, happiness = 0.69, enjoyment = 0.53), contributing 15%, and (iii) stress-related difficulties (insomnia = 0.65, stress = 0.61, academic difficulties = 0.45), accounting for 13%. The model demonstrated strong reliability (CFI = 0.975, TLI = 0.960, RMSEA = 0.066). Machine learning models detected regional differences in mental health. The stacking model (SM) achieved the highest predictive performance (accuracy = 98.0%, sensitivity = 88.0%, specificity = 100%). Other models performed moderately, including ensemble standardize (ESS) (accuracy = 74.3%) and voting classifier (VC) (accuracy = 70.7%). Dhaka students displayed higher concentration (1.57), resilience (1.75), and happiness (1.81), while Chattogram students reported greater enjoyment (1.83). Rajshahi (stress = 2.00) and Rangpur (insecurity = 3.00) students exhibited higher risks. Model accuracy peaked in Dhaka (98.0%) and dropped in Rangpur (74.2%). The stacking model prioritized accuracy over interpretability, highlighting the need for statistical models to enhance explanatory power. Findings emphasize targeted mental health interventions for students in rural areas to improve well-being in LMICs.