Deep E-Learning RecommendNet: An Acute E-Learning Recommendation System with Meta-Heuristic-Based Hybrid Deep Learning Architecture

Information and communication technology requires an adaptive learning concept for improving the outcomes of students based on their preferences. E-learning recommendation system with hybrid deep learning is developed for providing best recommending reviews for right users. Here, the different reviews from various E-learning platforms were taken as the input text data. These data are given into the preprocessing by performing tokenization, stemming and stop words removal to obtain the preprocessed text data. Then, the term frequency-inverse document frequency (TF-IDF) and glove embedding techniques are utilized for extracting text features. Further, the weighted feature selection is performed, in which the weight optimization and the selection of optimal features are done using the adaptive cat and mouse based optimizer (ACMO) strategy. The weighted features are further involved for the recommendation stage by developing a hybrid deep learning network named DELRNet that includes the combination of recurrent neural network and autoencoder for recommending the best E-learning platform for the users. The parameters in deep architectures are tuned with help of ACMO strategy to improve the effectiveness of the recommendation model. The experiments are carried out on the developed E-learning recommendation system to elevate the efficiency by comparing with other comparative approaches using different measures.