No abstract

No abstract

No abstract

Objectives:To perform aspect extraction for opinion mining of learner comments in the online teaching-learning scenario. Methods : A machine learning model is developed for aspect extraction. The authors collected the dataset consisting of around 5000 learner comments from Coursera and performed analysis on the dataset. To validate the results, the standard SemEval2014 dataset is used. Findings: For both the contextualised and non contextualised word embeddings, the authors compiled the results of their model for two different datasets, viz., the SemEval 2014 dataset and the collected dataset. The proposed model provided an accuracy of 92.78%. Novelty: For the first time in the literature, double embedding is used for aspect extraction. Using domain-specific embedding provided good results, but the proposed model (using double embedding) is far superior in accuracy.