Students Performance Measurement and Prediction Based on Academic Features Through the Machine Learning

Abstract

The education sector plays a vital role in achieving long-term economic and national development. During the last decades, due to the availability of resources and technology, efficient and higher standards in education have become easier to attain. The amount of data in educational institutions is growing rapidly. Through the data mining and machine learning methodologies, it has become easier to look into data from a different perspective and extract various information from the data. In our research, we use various algorithms to find out the correlation between features and predict students’ performance using their academic records. We try to find out the factors which affect and influence the performance of students. We implemented both unsupervised and supervised learning algorithms in our research. K-mean clustering has been used as an unsupervised learning method to group the students and find out the dependencies between the features. For prediction purposes, we use classification techniques like KNN and Linear Regression to predict students’ performance. Our research not only aims into finding useful information but also provides insight into students’ preferred teaching methods, potentiality, and performance. This information can guide the students for their future and guide them to their preferred fields according to their skill sets.