Published Paper Details:

Automatic Student overall performance prediction is a vital job due to the giant extent of statistics in instructional databases. This job is being addressed by means of academic records mining (ARM). ARM increase strategies for discovering statistics that is derived from academic environment. These techniques are used for grasp scholar and their gaining knowledge of environment. The academic establishments are frequently curious that how many college students will be pass/fail for indispensable arrangements. In preceding studies, it has been found that many researchers have intension on the resolution of gorgeous algorithm for simply classification and ignores the options of the issues which comes all through facts mining phases such as statistics excessive dimensionality, category imbalance and classification error etc. Such sorts of troubles decreased the accuracy of the model. Several universal classification algorithms are utilized in this area however this paper proposed a pupil overall performance prediction mannequin based totally on supervised studying choice tree classifier. In addition, an ensemble approach is utilized to enhance the overall performance of the classifier. Ensemble techniques strategy is designed to clear up classification, prediction problems. This find out about proves the significance of statistics preprocessing and algorithms fine-tuning duties to get to the bottom of the facts high-quality issues. The experimental dataset used is acquired from UCI Machine Learning Repository. Supervised mastering algorithms Decision Tree (J48) are employed in this find out about for experimental purposes. The effects confirmed that J48 performed absolute best accuracy 95.78% amongst others.