Predicting students' academic performance: Levy search of cuckoo-based hybrid classification

Abstract

Educational Data Mining (EDM) exists as a novel trend in the Knowledge Discovery in Databases (KDD) and Data Mining (DM) field that concerns in mining valuable patterns and finding out practical knowledge from the educational systems. However, evaluating the educational performance of students is challenging as their academic performance pivots on varied constraints. Hence, this paper intends to predict the educational performance of students based on socio-demographic information. To attain this, performance prediction architecture is introduced with two modules. One module is for handling the big data via MapReduce (MR) framework, whereas the second module is an intelligent module that predicts the performance of the students using intelligent data processing stages. Here, the hybridisation of classifiers like Support Vector Machine (SVM) and Deep Belief Network (DBN) is adopted to get better results. In DBN, Levy Search of Cuckoo (LC) algorithm is adopted for weight computation. Hence, the proposed prediction model SVM-LCDBN is proposed that makes deep connection with the hybrid classifier to attain more accurate output. Moreover, the adopted scheme is compared with conventional algorithms, and the results are attained.