A Naïve Based approach of Model Pruned trees on Learner’s Response

Appraisal and feedback have a strong positive influence on teachers and their work. Teachers report that it increases their job satisfaction and, to some degree, their job security, and it significantly increases their development as teachers. Student’s appraisal towards a teacher plays a vital role in building a very good teaching-learning environment in an educational institution. The evaluation report of the student helps the stakeholders to retain qualified teachers for the course. It will also help the teacher to understand the need of the student and the course. Therefore it becomes necessary to evaluate the teacher using appropriate tool to improve the quality of the education. Teacher evaluation can be measured based on the technical knowledge, communication skills, clarity, attitude towards the student etc. Regression trees can be considered as a tool to analyze the teacher appraisal scores. Two regression trees namely the REP tree and M5P algorithms are applied on the data set to bring out new knowledge from it. The algorithms have identified Parameter A as an important factor in teacher’s appraisal. Pruning has been taken as parameter to find the accuracy of the algorithm. The performance of the algorithm is measured using the mean absolute error and the time taken by the algorithms to derive the regression tree. The REP tree algorithm performs better than the M5P algorithm in terms of accuracy as well as the performance.