• Sujith Jayaprakash
• Jaiganesh V.

Higher education institutions across the globe are tirelessly making efforts to improvise the student performance and curb down the attrition rates. Academic institutions hold a massive amount of information like students personal details, classroom activities, academic achievements, etc., Thus obtaining insights from this data is highly imperative to take crucial decisions for education institutions. Embracing technology is the only solution to address these challenges and this in turn attributed to the usage of Education Data Mining. In this research work, we made a conscious effort to predict the performance of students using various attributes like socioeconomic, academic and demographic details. To overcome the class imbalance problem, we applied the Synthetic Minority Over-sampling Technique (SMOTE) technique. We conducted Pearson's Chi-Squared Test to identify important attributes and Ensemble methods were used to build the prediction model. The results showed that Random Forest performed better than the Bagging and Boosting models.

SMOTE, Chi-Square Test, Random Forest, Bootstrap Aggregation, AdaBoost

"Improvising multinomial Classification Accuracy of the model using Feature Selection and Class Imbalance algorithms ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.6, Issue 4, page no.41-46, April-2019, Available :http://www.jetir.org/papers/JETIRBI06008.pdf

"Improvising multinomial Classification Accuracy of the model using Feature Selection and Class Imbalance algorithms ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.6, Issue 4, page no. pp41-46, April-2019, Available at : http://www.jetir.org/papers/JETIRBI06008.pdf