الفهرس 
chapter 1 : IntroductionOnly 14 pages are availabe for public view
 |  | from | 100 |  |  |
| | | from | 100 | | |
Abstract
The process of solving the problems related to e-learning is carried out in several stages and is in order (developing an integrated website including a system of recommendation - data collection and processing - applying methods of data mining to data collected). The website’s recommendation system ensures that the site’s goals are met and its problems addressed (poor educational content on the site - carelessness by the instructor - monitoring the progress of the learning process). Data pre-processing by determining useful attribute in the prediction process before applying data mining techniques. After that, data mining techniques for best results.This Thesis aims to improve students’ performance in secondary e-learning systems by employing data mining (DM) models. Data is collected using online school tests, reports and quizzes. The data mining algorithms which use this data are support vector machine (SVM), Decision tree, M5-Rules and Linear Regression. Evaluate the result of the decision tree and SVM algorithms by accuracy ratio. On the other hand, the result of M5-Rules and Linear Regression is evaluated by error percentage. This thesis applies SVM with accuracy 89%, Decision tree with accuracy 89%, M5-Rules with RMS error equal to 1.4621 and Linear Regression with RMS error equal to 2.0017 for Third grade Secondary, 3.0089 for First grade secondary and 3.6057 for Second grade secondary. Once getting both first and second grades, the results presented a high predictive accuracy. Not only past student’s evaluations affected in their academic achievement, but also other factors like father’s and mother’s jobs and absences. Briefly, student performance can be improved depending on predictive results and enhancing school systems.