Predicting Educational Performance using Data Mining Techniques

Document Type : Original Article

Authors

1 Faculty Member of Industrial Engineering Group, University of Gonabad, Gonabad, Iran

2 University of Gonabad

Abstract

Identifying factors that affects students’ performance can help universities and higher education institutions in improving the quality of education. The knowledge acquired from educational data can help decision makers in understanding students’ behaviors and improving students’ performance. In this research, using the data collected from the students, some models for predicting their educational performance (in terms of their average score) are proposed. To develop these models, first, the affecting factors are identified using the best subsets regression and genetic algorithm. Then, using the identified factors, the average score of the students is predicted by the decision three and neural network methods. The results revealed that the predictions from the decision tree algorithm based on the factors obtained from the genetic algorithm were more accurate than from the others. Beside the real data collected from the students, the use of different data mining techniques provides rich information, which made the decisions more sensible.

Keywords

Main Subjects