SMOTE based Ensemble model for educational data mining

Abstract

Abstract Data mining in the classroom is one of the prominent fields which involves data mining concepts, statistical analysis, and machine learning concepts which all gets applied on the educational data. These EDM processed data are widely used for analysing the various aspects of the business and process model. Existing and conventional process model involves the usage of conventional statistical techniques to process the data which in-turn needs a lot of manual interventional for data modelling and pre-processing. To address the above-mentioned issues this paper proposes a novel technique which combines the machine learning model along with the statistical approaches. This machine learning combination involves the ensembling different classifiers such as Decision tree, logistic regression, K nearest Neighbour, Random Forest, multiplayer perceptron etc. The data which utilized in the experimentation is highly imbalanced due to the limited data availability. Hence the above claimed technique is combined with the universally benchmarked model called synthetic minority oversampling technique (SMOTE) to address the problem of class imbalance. Further, the performance evaluation is also statistically performed in order to prove the efficacy of the proposed technique.