Analysis of Students Emotion for Twitter Data using Naïve Bayes and Non Linear Support Vector Machine Approachs

Abstract

Students' informal discussions on social media (e.g Twitter, Facebook) shed light into their educational understandings- opinions, feelings, and concerns about the knowledge process. Data from such surroundings can provide valuable knowledge about students learning. Examining such data, however can be challenging. The difficulty of students' experiences reflected from social media content requires human analysis. However, the growing scale of data demands spontaneous data analysis techniques. The posts of engineering students' on twitter is focused to understand issues and problems in their educational experiences. Analysis on samples taken from tweets related to engineering students' college life is conducted. The proposed work is to explore engineering students informal conversations on Twitter in order to understand issues and problems students encounter in their learning experiences. The encounter problems of engineering students from tweets such as heavy study load, lack of social engagement and sleep deprivation are considered as labels. To classify tweets reflecting students' problems multi-label classification algorithms is implemented. Non Linear Support Vector Machine, Naïve Bayes and Linear Support Vector Machine methods are used as multilabel classifiers which are implemented and compared in terms of accuracy. Non Linear SVM has shown more accuracy than Naïve Bayes classifier and linear Support Vector Machine classifier. The algorithms are used to train a detector of student problems from tweets.
DOI: 10.17762/ijritcc2321-8169.1505151