Comparative Analysis of Classification Algorithms for Student Performance

Author(s):

Veena N , B M S Institute of Technology and Management; Guruprasad S, B M S Institute of Technology and Management

Keywords:

Naïve Bayes, Support Vector Machine(SMO), K-nearest neighbors(IBK), Classification Algorithms, Student Performance

Abstract:

This paper aims to present a model for the students performance. The predictive model was developed based on students performance in the second semester. Classification techniques from Data mining were applied to develop the models like Naïve Bayes, Support Vector Machine (SMO) and K-nearest neighbors (IBK). Comparative analysis is conducted on the three selected algorithms to find the best classification model. Moreover, this research also aims to find out the most influential subjects’ grades on their study duration. Courses, gender, and grades that serve as the independent parameters to predict the dependent parameter. The resulting models of the three algorithms show no significant difference between Naïve Bayes and SVM performances, while K-NN has the highest performance. Basic subject’s grades found to be the most influence parameter to the students’ study duration, followed by general subjects, grades, gender, and major subjects grades parameters.

Other Details:

Manuscript Id	:	IJSTEV4I2035
Published in	:	Volume : 4, Issue : 2
Publication Date	:	01/09/2017
Page(s)	:	81-85

Download Article