Towards Optimization of Math Pedagogical Techniques and Behaviors Using Decision and Ensemble Trees

Professional development programs for educators are thought to be generally ineffective and inefficient in raising student enjoyment and performance. As a result, new educators often express that the first years in their career are challenging and lacking in appropriate and effective support. New educators may frequently experience stress and anxiety from those factors as well as the inherent challenges of teaching young learners. The objectives of this paper and further research are to derive the most effective teaching techniques and behaviors using decision and ensemble tree methods. Six different models were trained from data collected on Egyptian preparatory language school mathematics teachers and their respective students. Two decision trees, two random forests and two gradient boosted trees are included, one set for classification tasks and the other for regression. The Student Evaluation Index (SEI) employed and generated by the student survey responses in this paper are used to derive the importance rankings of teaching techniques and behaviors. The tree structures can also be used to help teachers create improvement programs that will increase the performance and enjoyment of their students. The results of SEI prediction utilizing our tree methods are found to be critically dependent on techniques like Visualization and the use of native Language. It stands to reason that preparatory language school teachers in Egypt should emphasize the use of visual aids while instructing students in their native language. Additional attributes of a popular or effective teachers include inspirational behavior and the usage of practice problem sets such as homework.