FEEDBACK OF STUDENTS ABOUT TEACHING LEARNING PROCESS AND INFRASTRUCTURE FACILITIES IN AN EDUCATIONAL INSTITUTION USING SENTIMENT ANALYSIS

Author (s) : R.SARALADEVI, S.ALAMELU

---

Abstract :
The process of analyzing the past behavior from entities is feedbacks which act as a key role for accomplishing expected results to current and future behaviors. Educational institutions have attempted to collect feedback from students to study their sentiment towards courses and facilitates provided by the institution for improving the college environment. Based on this purpose, the student‟s sentiments have found out and the model of sentiment analysis is developed from the piece of text. In present scenario, grading technique is used for feedback which does not reveal the true sentiment of students but there should be the chance to the students by providing the textual feedback to highlight the definite aspects. The method of finding and interpreting the student input feedback is important by using Natural Language Processing (NLP) which is otherwise known as Sentiment Analysis (SA). Nevertheless the broad use and success of some approaches, it is difficult to find an enhanced technique to define the polarity of a text details. This paper develops the knowledge of evaluating the performance of different classifiers in ML, namely Logistics Regression (LR), Support Vector Machine (SVM) and Naive Bayes (NB) to demonstrate their efficiency in the student reviews using sentiment mining. Feedback from the student has been collected through Google form and this research work had nearly 500 student‟s feedback reviews. It is utilized for the analysis based on related categories namely Course content, lab work, extracurricular activities, library facilities and examination. The performance evaluation of classifier can be measured in terms of the accuracy. Based on the results it is observed that LR has produced 90% with better accuracy than other classification algorithm.

Keywords : Feedback, sentiment analysis, students, education

Conference Date : 13/11/2020

Pages : 198

Paper ID : 20198