Student Placement Prediction Using Genetic Algorithm for Education Data Mining

Data mining has a key application domain termed as education data mining (EDM). The aim of EDM is to provide novel, useful and previously unknown patterns to academia. Student placement prediction, student drop out prediction, student group formation, students’ performance prediction and student failure analysis are few challenging problems in field of education data mining. For education domain stakeholders, student placement is one of the important parameter. Placement of students indicates major educational systems’ success of any institute. It is also a prime indicator of institute’s performance. Prediction of student placement help academic organization to analyze and understand the prime factors based on which, it can take prior actions. Machine learning algorithms provide excellent models which provides near accurate predictions. However, dimensionality reduction is an important step of pre-processing which should be carried out efficiently for useful outcomes. As dimensionality reduction is a NP-hard problem, evolutionary algorithms are candidate algorithms to apply. In this paper, genetic algorithm is used to perform student placement prediction. Paper demonstrates application and comparison of results among machine learning algorithms with and without use of genetic algorithm as a dimensionality reduction technique.