Learners’ Satisfaction Analysis Using Machine Learning Approaches

In this competitive world, the Universities have the challenge to genuinely analyze their performance with respect to teaching-learning process. The teacher and students should be answerable to each other. To analyze the teaching- learning performance, the feedback is very basic and essential tool. Here we present student feedback analysis concerning the instructor or educator using machine learning algorithms. In this paper, first, we grouped the feedback data from the University students to get a useful pattern with the help of clustering algorithms like K-means and EM (Expectation Maximization) and chosen the best one. After finding the clusters from feedback dataset, we have assigned three categories as, satisfactory, neutral, and dissatisfactory and used them as class labels for classification purpose. We have applied Naive Bayes, Multilayer Perceptron Neural Network, Random Forest (RF) and Support Vector Machine (SVM) classifier and found that Naive Bayes got the highest accuracy, precision, and recall values as compare to the other classifiers. The results obtained here indicate the satisfaction level of students with a particular instructor is less positive as compared to other instructors.