An Intelligent Hybrid Approach for Predicting the Academic Performance of Students using Genetic Algorithms and Neuro Fuzzy System

Higher education institutions aim to offer high quality education to its students, by monitoring and analyzing the performance of the students to provide appropriate means to the students with low performance to meet their needs in early stages during their academic career. The availability of large volume of data in educational databases has made the prediction of the students' academic performance a challenging and quite important task. To overcome this problem, utilizing soft computing is one of the most effective and efficient methods for solving different kinds of problems, especially predicting the student's academic performance. Among many soft computing approaches, the neuron fuzzy inference models which are considered the most appropriate approach. This paper presents an intelligent hybrid genetic neuro-fuzzy inference system (HGANFIS) for student's academic performance prediction. HGANFIS intelligently integrates the learning and reasoning capabilities of the Adaptive neuro fuzzy inference system (ANFIS) with the powerful optimization of the genetic Algorithm (GA). The proposed approach HGANFIS utilizes the previous exam results of the students as input variables to ANFIS, and incorporates the GA into ANFIS to acquire the optimal ANFIS parameters in order to obtain accurate prediction of student's performance. The experimental results were appealing and showed that the proposed HGANIS obtained highest student's performance prediction accuracy when compared with other approaches.