Sentiment Analysis for E-Learning Counting on Neuro-Fuzzy and Fuzzy Ontology Classification

Abstract

Nowadays, the utmost effective rising approach in learning procedure is E-Learning. To be specific, the combined learning is meant to be a worthwhile approach for assisting and comprehending schoolchildren and their acquisition matters. Because of E-Learning platforms and their collaborative tools, schoolchildren might communicate with other schoolchildren and share doubts on definite course or with teaching staff. When we try to enhance E-Learning, it is exceptionally worthwhile to have enlightenment about the users’ sentiments available. On one side, this information might be utilized by adaptational E-Learning systems for the drive of assisting personalized learning, by considering the user’s emotional state when recommending him/her the utmost suitable activities to be undertaken at each time. Contrarily, the schoolchildren’ sentiments concerning a course might serve as feedback for teaching staff, specifically in reference to online learning, where face-to-face interrelationship has been considered to be a lesser great deal of frequency. The worth of this study in the perspective of E-Learning, both for teaching staff and for adaptive systems, is looked at too. Adoption of an innovative means for sentiment analysis through Facebook is inspected, beginning with messages written by schoolchildren to mine enlightenment about the schoolchildren’s sentiment polarity (positive, neutral or negative), as conveyed in the messages written by them. We have used this methodology in SentBuk, a Facebook application also put forward in this study. The classification methodology implemented follows a hybrid approach: It combines neuro-fuzzy algorithm and fuzzy ontology. Results acquired through this model show that it is feasible to carry out sentiment analysis in Facebook within height accuracy.