Performance analysis of educational data mining framework for higher education using machine learning

Abstract

Predicting student performance is one of the most significant themes newlinein learning contexts because it aids in the creation of successful mechanisms to newlineimprove academic results and prevent dropout. Instructors can more correctly newlinedistribute newlineresources and teaching using effective performance prediction tools whereas, newlineidentifying the factors affecting student s performance in higher education, newlineparticularly through the use of predictive machine learning, is still an ongoing newlineresearch. newlineCurrent eras of students at higher education are lethargic in appearing newlinefor exams as well as in maintaining the minimum marks required to get qualified newlinefor the entire course. According to NEP2020, all graduate students should be newlinequalified at the desired educational level. Until the students received the newlineunqualified grade, they did not know their actual academic status. Even if the newlineoral instructions are given to the students, they are all ignored in a very brief newlinetime. newlineAt the time they realize and work on it, they can be skilled in second or other newlineattempt which also affects the percentage of institutional success. newlineThe field of educational data mining and specially, the warning newlinesystem based on prediction faced many issues and it is all time research field. newlineThe problem of slow learners in academic institutions is vital since an indication newlineshould be triggered before the final assessment. Even though, there are few newlineworks found previously based using data mining feature selection, the reality is newlinethere is no best set of input variables for a problem that needs dynamic solution. newlineIt should be discovered what works best for the problem chosen through newlinesystematic experimentation. Here comes a need for a dynamic technique to find newlinethe best playing features. There is no such case where a same model can provide newlinethe best output for a dynamic problem all the time. newline