Özet:

Abstract In recent years, academics from many related study fields worldwide have begun to focus on educational data mining (EDM). To help academic planners in higher education institutions make better decisions, suggestions can be made using the information gained from the EDM. Various prediction models have been put out in the literature to forecast student performance. This paper suggests a distributed cluster-based architecture (CDA) for predicting student performance. The proposed CDA indicates clustering via water wave optimization based on K-means cluster and deep neural network (WWO-KMC-DNN), feature extraction using Multi-Linear Discriminant Analysis (M-LDA), and feature fusion using a Bayesian network. In the suggested design, the WWO algorithm is used to determine the DNN ideal weights. Accuracy, prediction rate, mean square error, and root mean square error is monitored in a real-time database to evaluate the proposed task. Using the MSE and RMSE values from the data, the proposed WWO-KMC-DNN model outperforms other models.

Anahtar Kelimeler: