In recent times, quality of teaching and students’ participation is monitored in most of the universities and colleges to improve academic performance. Outcome-based education system also requires experiential learning. Student opinion is one of the powerful mechanisms to evaluate academic activity and quality in education system. It helps to decide on corrective measures towards various entities of teaching and learning process. Many research studies have been carried out in classifying the sentiment polarities of opinions. In educational domain, very limited study is carried out in fetching the key aspects or entities from the students’ review. Mining the opinions in a deeper level to fetch the specific aspects confined to the topic will bring productive results. In this paper, aspect-based sentiment analysis is carried out at sentence level to find the students’ satisfaction with reference to the online courses using machine learning algorithm. The proposed system has attained improved accuracy than the existing model. From this study, the specific aspects are obtained using both unsupervised and semi-supervised LDA algorithms. Students’ satisfaction with specific aspects of the online courses also examined.