University Students Result Analysis and Prediction System by Decision Tree Algorithm

Abstract

The main assets of universities are students. The performance of students plays a vital role in producing excellent graduate students who will be the future viable leader and manpower in charge of a country's financial and societal progress. The purpose of this research is to develop a "University Students Result Analysis and Prediction System" that can help the students to predict their results and to identify their lacking so that they can put concentration to overcome these lacking and get better outcomes in the upcoming semesters. The prediction system can help not only the current students but also the upcoming students to find out exactly what they should do so that students can avoid poor achievement that will help to increase their academic results and other skills. To train the system, we collected data from the university student's database and directly from students by survey using Google form containing information, such as gender, extracurricular activities, no of tuition, programming skills, class test mark, assignment mark, attendance, and previous semester Grade Point Average (GPA), where the main aim is to relate to student performances and Cumulative GPA (CGPA). We use Weka tools to train the system and to develop the decision tree. In decision tree, the acquired knowledge can be expressed in a readable form and produced classification rules that are easy to understand than other classification techniques. These rules used to develop a web-based system that can predict the grade points of students from their previous records. Moreover, the system notifies students' lack and gives suggestions to improve their results. Finally, we compared the performance of three (J48, REPTree, and Hoeffding Tree) different decision tree algorithms, and comparative analysis shows that for our system, the J48 algorithm achieves the highest accuracy.

Deep Biased Matrix Factorization for Student Performance Prediction

“…In the field of educational data mining, many studies have explored the use of decision tree algorithms for predicting student performance. One such study, conducted by the author of [13], compared three different decision tree algorithms and found that J48 was the most effective for classifying and predicting student actions. The study also noted that the decision tree graphs' structure was influenced by the number of input attributes and the end class attributes.…”

An Ensemble Modeling Approach to Enhance Grade Prediction in Academic Engineering Programming Courses

“…However, it should be noted that this study employed a more limited dataset, resulting in decreased accuracy. In [13], researchers employed three distinct decision tree algorithms-Hoeffding Tree, J48, and REPTree-to predict university results and aid students in improving their performance by identifying areas of weakness. The dataset for university students encompassed a wide range of information, including assignment marks, attendance, class test marks, extracurricular activities, gender, tuition frequency, programming skills, and preceding semester Grade Point Averages (GPA).…”

Model for Prediction of Student Grades using Data Mining Algorithms

“…The purpose of the proposed project was to create a more efficient framework that significantly improved student performance. Reference [12] used WEKA (Waikato Environment for Information Analysis) as a testing tool for data classification. Working on this paper was very difficult with tree size, entropy, and other factors.…”