Pragmatic Actionable Knowledge Mining for School Student’s Dropout Prediction and Prevention

Janapati Naga Muneiah and Ch D.V. Subba Rao

School students’ dropout rate is still growing in many countries. Some students are permanently stopping going to school due to various reasons and this has been evolved as a big problem in the education system. Hence, it is necessary to detect the risk of dropout in advance and apply the required actions to avoid dropout. In our research, we introduce an efficient method that is based on a decision tree that can predict the probability that an existing student dropout from the school and also automatically suggests actions to prevent the school student from dropping out. Moreover, the proposed method tries to convert the student who is having a mediocre degree probability of Non-dropout class as a higher probability of Non-dropout class. Our method also suggests various sets of actions and the number of probable dropout students who can be changed as the Non-dropout students. The proposed machine learning based method uses the probability estimation decision tree to predict the student’s dropout nature. It adopts efficient data structures and employs a novel approach for suggesting the actions. Empirical tests are conducted to evaluate the efficiency of the proposed method and demonstrated that the proposed method outperforms the state-of-the-art methods.

Volume 12 | Issue 3

Pages: 21-29

DOI: 10.5373/JARDCS/V12I3/20201163