Title: A hybrid approach for deep belief networks and whale optimisation algorithm to perform sentiment analysis for MOOC courses
Authors: Jayakumar Sadhasivam; Ramesh Babu Kalivaradhan
Addresses: School of Information Technology and Engineering, Vellore Institute of Technology, Vellore, India ' School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, India
Abstract: Sentiment classification has won significant attention presently, as it provides the way for automatic analysis of people's reviews to extract user information regarding a product/service. One of the widely used techniques is polarity classification that determines the polarity of the texts in the opinion. Accordingly, this paper presents a technique for sentiment classification of online course reviews using a novel classifier, WDBN. In the proposed technique, the input course review data is pre-processed, and important features are extracted from the data using Emotion-SentiWordNet-based feature extraction process. For the classification of sentiments in the feature extracted data, WDBN is introduced by combining DBN and WOA such that the weights of the network layers are selected optimally. The proposed WDBN classifier is experimented using a publicly available online course review dataset and the performance of the classifier is evaluated using three metrics, such as sensitivity, specificity, and accuracy.
Keywords: deep belief networks; DBN; MOOC; whale optimisation algorithm; WOA; whale-based deep belief network; WDBN; neural-network.
DOI: 10.1504/IJAIP.2023.132372
International Journal of Advanced Intelligence Paradigms, 2023 Vol.25 No.3/4, pp.280 - 296
Received: 29 Nov 2017
Accepted: 15 Feb 2018
Published online: 19 Jul 2023 *