Predicting student's performance using CART approach in data science

In technical graduate courses like engineering, it is very important that students are monitored during their first year. It is important for all technical graduates to have a good programming insight. Still, a student shying away from the core subjects like programming is common phenomena in engineering colleges nowadays. It is thus necessary to keep track of students' performance during the first year of the course especially. During semester beginning, poor performance further creates disinterest in students and hence the end result is low whether it is on semester card or it is their knowledge level. It gives a scope for us to apply data science techniques to analyze and predict a student's performance. To address this problem we propose a system which will make use of the decision tree approach to predict a student's performance. Based on student's current performance and some measurable past attributes the end result can be predicted to classify them among good or bad performers. It will thus enable a faculty to pay attention to the weak students and plan sessions for them accordingly. A student's interest in programming or other core subjects can be monitored and encouraged as needed.