Learner performance prediction in the e-learning platform using the optimized deep long short-term memory classifier

Abstract

The e-learning platform gains significant attraction in the current scenario due to the outbreak of the epidemic COVID-19 as e-learning ensures the students continue their studies in the safest environment while maintaining the educational standard. The performance prediction is one of the significant tasks to be carried out in the e-learning platform to sort out the students who require immediate attention to enhance their grades before the final assessment. This paper proposes a prediction model that effectively predicts the learners’ performance in the e-khool learning management system (e-khool LMS) based on the proposed wolf-swarm optimization dependent Deep Long Short-term (wolf-swarm optimization-based Deep-LSTM) approach. The optimization algorithm tunes the optimal weights of the Deep-LSTM classifier, which inherits the hybrid characteristics of the traitors and particles. Initially, the learner data from the e-khool database is employed for classification based on the proposed wolf-swarm optimization dependent Deep-LSTM classifier. The effectiveness of the proposed prediction model is analyzed in terms of MSE and RMSE with the value of 5.93 and 2.426, respectively.