Abstract

In recent era of Digitization, large amount of computer generated data is accumulated on servers. However, these gathered data is useful only when interesting, novel and applicable knowledge is generated out of it. In the field of education, digitized data is generated by online academic activities. Data mining is discussed with every aspect of society however its use in academia is at infancy. Data mining is used to find novel, interesting and useful knowledge out of data which directs to actionable patterns on which academicians could work to enhance the productivity of academic activities. Education data mining is focused to use educational data for knowledge discovery to attain valuable insights in education domain. In this paper, an important task of prediction of institute graduation rate is addressed. Two novel approaches are proposed in the paper for effective graduation rate prediction. The first approach is genetic algorithm assisted regression model. The second approach investigates and uses various filter methods to further enhance the results in terms of time and number of features. Three regression models – multiple linear regression, decision tree regression and support vector regression are considered for experiments and comparative results are produced. The proposed methods provide better institute graduation rate prediction.