Developing a Predictive Model to Improve Student Outcome Using Big Data Analytics and Machine Learning Techniques

Abstract

Today’s student outcomes have become one of the most challenging problems for academic institutions. To improve student outcomes one should understand the non-trivial reason behind the loss of students and to be successful, one should accurately identify those that are at risk. Nowadays students are not achieving the outcomes they expected in academic institutions. So far, many studies have been conducted to comprehend the application of big data in different fields for various purposes. However, data analytics is still lacking in big data in education. Hence, this research looked into the use and impact of big data analytics in public universities and secondary schools in Ethiopia. Big data provides unprecedentedly rich information for machine learning algorithms to extract underlying patterns and build predictive models. The main objective of this study is to develop a predictive model using big data Analytics and machine learning techniques to improve student outcomes. These data sets are in tabular format, where each row represents a student and each column, or variable, contains certain information about a student. Data was obtained and analyzed using various tools/software such as Excel for data preprocessing and Python for data analysis. The method used was experimental research using data collected from universities and high schools. For the university models the predictions are checked on machine learning algorithms such as random forest with accuracy values of 99%, support vector machines with accuracy values of 98% and naïve bayes with accuracy values of 87% and for the high schools model k-nearest neighbored with accuracy values of 95%, random forest with accuracy values of 99% and support vector machines with accuracy values of 98%, and the results are presented & discussed. In this study, the determination of student status using machine learning techniques was only done for 2 public universities and 8 secondary schools but we proposed other researchers to explore the other private or public university and schools trends of student status.