Aspect-Based Opinion Mining of Students' Reviews on Online Courses
Zenun Kastrati, Blend Arifaj, Arianit Lubishtani, Fitim Gashi, Engjëll Nishliu
Abstract
It is critical for higher education institutions to work on improvement of their teaching and learning strategy by examining feedback of students. Analyzing these feedbacks typically requires manual interventions which are not only labor intensive but prone to errors as well. Therefore, automatic models and techniques are needed to handle textual feedback efficiently. To this end, we propose a model for aspect-based opinion mining of comments of students that are posted in online learning platforms. The model aims to predict some of the key aspects related to an online course from students' reviews and then assess the attitude of students toward these commented aspects. The proposed model is tested on a large-scale real-world dataset which is collected for this purpose. The dataset consists of more than 21 thousand manually annotated students' reviews that are collected from Coursera. Conventional machine learning algorithms and deep learning techniques are used for prediction of the aspect categories and the aspect sentiment classification as well. The obtained results with respect to precision, recall, and F1 score are very promising.