تنقيب بيانات طلاب المرحلة الاساسية للتنبؤ بدرجاتهم فى مادة الرياضيات

Abstract

With the development of computer science, a lot of data became available in various disciplines and educational institutions and one of the systems that need to extract its data, which prompted the researcher to create a model for predicting students ’grades, and therefore the research problem centered on this study about students’ declining grades in mathematics according to students ’records in Basic schools, so there was a need to find a method by which to predict students' grades using the outcome dependent variable at an early stage. It aims to categorize students into two groups (pass and fail) to predict students' grades at an early stage. The researcher followed the descriptive and analytical approach to its relationship with the topic under study, as it collected a sample of 1500 samples in the year 2015-2018, dividing the sample into two groups. The first group represents the training group of 500 students, while the second group represents the test group of 1000 students. The researcher relied on the previously known scores of the students' results for the purpose of analyzing the variables that affect future changes in the decision tree algorithm using an orange data classification tool to predict the final grades of students based on their previous grades. The researcher used five tree algorithms to predict their results and then compare them with other algorithms: Naive Bayes and NN. Logistic Regression, SVM, and the best-performing decision tree was 99.7%.