Comparison of Machine Learning Techniques to Predict Academic Performance of Students

Many organizations use machine learning to analyze data and find significant hidden patterns in the data including healthcare, finance, online service provider, education institute, software companies etc. and based on getting rules from the pattern take the appropriate decisions in the favor of organization. This research paper has used four machine learning techniques to generate the models. This models are compared by various accuracy measured parameters to find the best suited model for the student's dataset. This paper has used various academic and demographic parameters of students to create the dataset. This research article includes logistic regression, decision tree, artificial neural network and Naive Bayes machine learning techniques. As accuracy measurement parameters on model this research article has used ROC index, Error Rate, F-measure, and accuracy. The data set is collected from sharing the drive sheet among students of various institute. As a result, found that ANN model is best suited and highest accurate model for this dataset so by applying this model institute got the highest accurate result and take the wise decision to improve the performance of students in academic.