Identify the variables that help explains the performance of the student in the Math subject with the support of the real-world data
In this dataset, all the variables were mostly of non-linear relationship, which makes Multiclass Decision forest to be a great choice. For the purpose of my analysis. I have removed activities, school, address, Medu, Fedu, Freetime, schools, guardian, reason. As we are not grouping the data based on the type of school, the name of the school will be of least significance. And other factors seem less significant because in score model results these features appear of least significance.