Evaluation of higher education institutions using aspect based sentiment analysis

Abstract

Demand for formal higher education programs among the younger generation in Sri Lanka, has grown over the past decade. The demand growth has fueled the opening up of many local and internationally affiliated institutes offering a diverse range of degree programs. The selection of the appropriate course from these institutes is challenging given the wide choice. In order to select the appropriate institute, students use the Internet for reviews and user comments, especially from social network sites like Facebook, Twitter and Google plus. This search, involves a cost in terms of time spent for reading the comments and processing whether the standing of the ratings for the program and the institution are appropriate. This task is challenging because of the difficulty to extract sentiment information from a massive set of online reviews. A solution is proposed, using an aspect based sentiment evaluation system that assesses institutions by considering the reviews provided, to overcome this problem. This concept is based on Natural Language Processing (NLP). A web based, automated application tool that retrieves review data from social media networks on the institution and the features of the program, analyzes the sentiment value and provides a rating has been developed.