Assessments of Student’s Adaptability Using Convoluted Geyser Bidirectional Long Short-Term Memory in Online Education

The rapid transition from traditional in-person education to online classrooms has highlighted the need to effectively assess student engagement in virtual learning environments supported by learning management systems. Despite this shift, there remains a significant gap in predictive models that generalize across diverse blended courses, disciplines, and student demographics. Addressing this gap is essential for improving the accuracy and efficiency of predicting student adaptability in online education. To address these issues, this research introduces a novel student adaptability learning model named Convoluted Geyser Bidirectional Long Short-Term Memory (CGBiLSTM) model, designed to predict student adaptability in online entrepreneurship education. This technique has its ability to capture complicated patterns and dependencies in sequential data, which is critical for accurately assessing student adaptability. Furthermore, incorporating the Geyser Optimization Algorithm (GOA) into CGBiLSTM improves the performance by optimizing the learning process and training capabilities, resulting in more accurate and dependable predictions. The CGBiLSTM technique achieves an accuracy of 98.94%, a precision of 99.03%, a recall of 98.71%, and an F1-score of 98.15% proving its efficacy in assessing student adaptability. The CGBiLSTM model, enhanced by the GOA, provides a highly accurate and reliable solution for predicting student adaptability in online education, making it a vital tool for educators in the evolving virtual learning landscape.