404
Journal of Education & Social Sciences - Volume 6, Issue 2 2018
By Noman Islam
10.20547/jess0621806204
Keywords: Faculty evaluation, machine learning, clustering, sentiment analysis, speech recognition, topic classification
In this paper, a three pronged solution to faculty evaluation is proposed. Almost in every university, faculty and course evaluations are filled by students after the completion of courses. Due to the large volume of such evaluations, it becomes very difficult for management to carefully analyze them. This paper proposes a framework based on machine learning techniques that can be adopted for effective evaluation of faculty. It uses k-means clustering to group the evaluations and points out the specific area on which management needs to work on with faculty. Along with the quantitative evaluation of faculty, students also provide feedback in the form of comments. The proposed solution performs sentiment analysis on those comments. If there is a high emotion (positive or negative) associated with comments, an email can be sent in real-time to higher management. Another important component of proposed solution is providing summary of the topics discussed in the lectures via transcribing their recorded lecture and then applying machine learning on transcripts.