LMS Content Evaluation System with Sentiment Analysis Using Lexicon-Based Approach

Abstract

The emergence of information technology used in all factors of our everyday lives has exponentially increased the amount of unstructured data. This huge quantity of records is a great source for finding and thus, may be used for extracting actionable information. In the academe, for instance, teachers and school administrators can adjust their approach to teaching/learning by getting feedback from students through a Learning Management System that can automatically analyze the semantic orientation of words and contextual polarity of these feedbacks - categorizing them into positive and negative. Identifying and classifying words expressed in the students' feedback about learning materials can provide structured information that can guide the teacher, impact its design and target the students' needs. This study implemented a lexicon-based strategy for automatic sentiment analysis using VADER as a model. Student feedbacks are extracted from an LMS developed to demonstrate the usability and effectiveness of the adopted approach; among other features of LMS that will help teachers improve its implementation. Results of the LMS sentiment analysis are compared to human-annotated sentiments to verify and validate the output, as well as, check its accuracy using Confusion Matrix. It aims to create a structured representation of student sentiments through LMS to help teachers improve the design of learning materials.