A Hybrid RF with BERT Model for Improved Eco-Anxiety Prediction among Medical Learners by Multi-Factor Data Fusion
Abstract
In the domain of healthcare, mental well-being analysis is more crucial one. The aspects that affect mental wellness are academics, and psychological factors over medical students. But growing occurrence of environmental impacts like floods, heatwave, air pollution affects the student’s anxiety levels, rises the eco-anxiety score. Furthermore, medical students under goes lots of academic stress, which affects their day-to-day life. Existing studies involves factors of academic and psychological, and failing to combine the major factors. Additionally, the past studies limited with synchronized dataset, and restricted multi-factor analysis for reliable prediction. To overcome these limitations, an integrated dataset with multi-factors, like environmental, academics, and psychological variables. The research is based on actual dataset by a proposed Hybrid RF with BERT model. The Climate Exposure Index (CEI) from climatic features, Academic Pressure Index (API) from academic features, GAD-7 and CES-D measure for mental health, are the essential factors of the dataset. The proposed model achieved excellent predictability by higher accuracy of 93%, and 0.93 of AUC for detecting the students at high-risk. The outcomes show the factors are depended on one another in predicting Eco-Anxiety level, emphasizing the need for early identification and adaptive mental wellness in medical educational framework.