International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 11 Issue: 05 | May 2024
www.irjet.net
p-ISSN: 2395-0072
Leveraging Machine Learning for Prognosticating Academic Performance Swetaben Vinodray Katariya1, Mayur Jani2 1CSE Department, Dr.Subhash University, Junagadh, Gujarat, India 2IT Department, Dr.Subhash University, Junagadh, Gujarat, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The study aims to predict students' educational
to high dropout rates, which lead to enormous loss and waste of resources. Studies reveal that the reduction in engineering programs outpaces that of all other fields of science and the arts [2].
achievements using machine learning algorithms, specifically the SVR (Support Vector Regression) methodology. Our goal is to determine students' final academic results to help reduce dropout rates and enhance the quality of education. By accurately predicting these outcomes, we hope to contribute to lowering the number of dropouts and improving educational standards. A recent investigation explored how machine learning algorithms can be used to create models that categorize different levels of grade point averages, academic retention rates, and degree completion outcomes. This study used data from students at a private university and found that the algorithms performed with high accuracy. Among various factors considered, learning methodologies stood out as the most significant predictor of grade point averages .
Another study employs three prediction models—support vector machines (SVM), neural networks (NN), and K nearest-neighbors (KNN)—to identify pupils who are at risk early on. Through a survey, 800 students' data were gathered for this study [3]. Another study investigates the development of an early warning system for identifying pupils in need of help through the use of machine learning (ML) algorithms. Using only the first evaluation of the course, the authors were able to identify students who could fail the course with a 95% accuracy rate. Furthermore, they had an 84% accuracy rate in identifying borderline students [4].
Key Words: Prediction, academic performance, machine learning algorithms, SVR
2. RELATED WORK
1.INTRODUCTION
A college or university degree is considered necessary in today's society for both responsible social engagement and economic advancement[5]. Nevertheless, obstacles that students face during their academic careers cause some of them to give up on their studies or take longer to get their degrees[6].
The field of machine learning has attracted a lot of interest lately from a variety of industries, including education. Algorithms for machine learning have been used to predict and evaluate students' academic performance, providing teachers with important information about how well their students are understanding the material and how effective their teaching strategies are.
According to [7], their investigation focused on two main aspects of undergraduate students' performance when using data mining (DM) techniques. Anticipating students' academic performance after a four-year curriculum was the main objective. Monitoring student progress and combining it with predicted data was another goal. Based on their performance levels, the students were divided into two groups: poor achievers and high performers [8]. The study underlined how crucial it is for teachers to closely monitor certain courses that have noteworthy strengths or shortcomings. With this targeted approach, prompt notifications, support for underachieving children, and opportunities for coaching and enrichment for high achievers are all made possible.
A strong education is built on a number of essential elements. Discipline, leadership, inspiration, academic qualifications, outside involvements, instructional strategies, academic supervision, creativity and research, character traits, and administration are all included in this. As a result, scholastic credentials and stellar performance are crucial for students' future opportunities. Algorithms for machine learning offer the chance to predict student outcomes in advance[1]. With the use of these algorithms, we are able to identify difficult students early on and offer them extra support and resources to meet their needs. By taking a proactive stance, we can provide customized, high-quality education that fulfills the needs of each individual student, which eventually lowers dropout rates and promotes higher enrollment.
In a similar vein, researchers [9] examined sixteen demographic variables, including age, gender, class attendance, internet accessibility, computer ownership, and course enrollment numbers, in order to forecast students' academic achievement. With the use of several machine
Academic institutions face major issues in educational settings and in evaluating and assessing their students due
© 2024, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 2135