Extract Information to Improve Teaching Effectiveness Using Machine Learning Methods

In traditional curriculum evaluations, educators often suggest improvements based on their impressions of teaching effectiveness and student achievement. However, this approach is not based on empirical evidence and can lead to subjective and potentially biased initiatives. To address this problem, our research introduces a data-driven approach to unearthing teaching reform strategies using the Random Forest classifier. Under this framework, students provide a comprehensive self-assessment score, as well as ratings in eight different areas. These scores are then used to calculate the achievement of each course objective, as outlined in the syllabus, which helps to identify areas that need enhancement and suggest targeted interventions. In addition, a threshold-based approach is used to classify the students' comprehensive assessment scores. These data, together with the self-assessment scores, make up the dataset of the Random Forest algorithm, which produces key features that influence classification. By examining these key characteristics, we can identify key teaching areas that need attention in future curricula and develop practical measures to be integrate into subsequent curricula.