Sensei - Teacher Rating and Comparing System Using Sentiment Analysis

Abstract

In the current education system students sometimes get chances to select teachers by their own. Specially in university level when selecting supervisors for undergraduate projects and tutoring. But choosing the right academic is not an easy task because students have so many options and they have to consider many facts. So in order to select the best academic most of the students go and look at academics‟ personal web pages, surf internet and check for the reviews and ask from seniors. But this is very time consuming and students can‟t get a clear idea analyzing all the resources they looked into. So the task is quite challenging for the students with the limited time they have with the current education system. As a solution for this problem, a system to rate academics using the reviews given by students who have previous experiences with the academic and using the social media profiles of the academic, is proposed. Also the solution is giving a rating for the qualities of the academics as well and the ability to compare academics with clear graphical representations. The implementation of the system is based on Natural Language Processing and sentiment analysis concepts. The final prototype of the system is implemented as a web application. The system achieved an accuracy level more than 80% and provides the user an effective, clear and understandable summary of the academics.