Performance Evaluation of Data Mining Classification in Educational System using Genetic Algorithm

Abstract

With the development in the field of Information Technology and Computer Science, high capacity of data appears in our lives. Data Mining helps us to find out useful information from large dataset. After retrieving information from large dataset, we have applied Genetic Algorithms to optimize the classified information. This paper provides a concise and representative review for classifying students in order to predict their performance on the basis of features extracted from the data logged in an Education System. We have discussed number of classifiers like 1-NN, K-NN, Naïve Bayes and Decision Tree (C4.5, C5.0, CART). The performance evaluation of these classifiers on students dataset is done using various attributes and we found that CART is the best data mining classification technique among the 6 classifiers when we use two classes and KNN is the best data mining classification technique among the 6 classifiers when we use three classes.