A fuzzy based enhancement on prism and J48 classifier prediction of student performance

A modified computed aided design of experiments (MCADEX) using Kullback-Leibler divergence and modified principal component analysis (MPCA) was proposed for selecting set of samples to improve the prediction of student's performance in Prism and J48 classifiers. The classification accuracy of prism and J48 is required to enhance further for classifying the students dataset which have complex related attributes. Hence, a fuzzy neuro prism (FNP) and fuzzy neuro J48 (FNJ48) are introduced for improving the classification accuracy. A fuzzy system is using fuzzy if then rules in obtaining knowledge from human experts can deal with imprecise problems. These rules are generating for describing the relationship among the input attribute space and classes. In fuzzification, Gaussian membership function is used. In this method, the weight value of each attribute is calculated using neural network. Fuzzy membership function parameters are optimized by using Cuckoo search algorithm. The attribute with maximum weight value and fuzzified value of features are used for constructing tree of prism and J48 classifiers. The experimental results show that the proposed approach is providing better results in terms of accuracy, true positive rate and true negative rate.