A Machine Learning Model for Predicting Academic Performance of Students Through Internet Usage

Internet is a powerful platform for students to develop the areas of interest and improve existing skills. Students in the age group of 18–25 are most frequent users. People in general, especially students, use the Internet mainly for research and educational purposes. The way in which people use the Internet meaning the websites browsed indicates the behavior seeking perspective. The outburst of Internet has aided students in many ways, but it also brings negative impact in academic performance. Study on people has also told that the balance maintained between students with study and Internet has a good impact in their academic performance too. The proposed work uses machine learning algorithms to quantify the relationship between performance in academics and behavior perspectives of a student in the usage of Internet and to bring out novel features that have a generalized value. Students are grouped according to academic performance and grades obtained with further processing by decision tree, support vector machines (SVM), and neural networks algorithms. Students’ Internet logs could be obtained and expose affluent information on students’ behavior. The proposed work has strong practical value for improving management skills of students’ in education with the university and college sector.