MOOC-LSTM: The LSTM Architecture for Sentiment Analysis on MOOCs Forum Posts

Abstract

The massive open online courses (MOOCs) have been among the foremost energizing improvements in e-learning environment in recent days. As the number of MOOCs resources on each domain growing greatly, there is a necessity of evaluating MOOCs. Discussion forums are the key resources for MOOCS evaluation. Sentiment analysis is the famous mechanism to identify the opinion of the students on every particular MOOC. Long short-term memory architecture is used to avoid the issue of long-term dependencies in the text. In this paper, we propose a sentiment analysis system contains a new LSTM architecture and Ax hyperparameter tuner that can jointly performs well with large text for sequential analysis and sentiment classification. Proposed system is trained on two different datasets from different platforms using optimal hyperparameters. Experimental results shown that the proposed system outperforms other machine learning models in terms of accuracy and working well with different domains.