Machine Learning Based Predicting Student’s Grade

Abstract

Many scholars have looked into student academic performance in monitoring and non-supervision supervised learning, and they have used a variety of data mining techniques. To gain significant guessing skill, neural networks often require a large number of observations. Due to the rising number of impoverished students with degrees, it is vital to develop a programme that aids in the reduction of this scourge as well as the incidence of recurrence due to poor performance or the necessity to drop out of school entirely in pursuit of their job. As a result, it is vital to understand each one’s benefits and drawbacks in order to identify which one works best and which ones should be chosen many times. The study’s goal is to create an Artificial Neutral Network-based student performance prediction system that uses student mathematical features to help the university select candidates (students) with a high prediction of admission success based on previous student academic records who will eventually earn a degree from the institution. The model was built using a set of input variables, including parental status. With parent’s data, there is about 80.84 percent accuracy, whereas without information about parent’s education, there is 81.37 percent, indicating that they are closely related. As a result, in the large scale of student’s data, these differences will be so minor, according to this machine learning prediction process. There are many things, which need to do predictive process to check the possibility of happening that thing, as the advancement is rapidly increasing.