Neural network and regression models for matriculation college

• Main Authors: Razali, Shahirawaty, Siraj, Fadzilah, Yusoff, Nooraini
• Format: Conference or Workshop Item
• Language: English
• Published: 2006
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://repo.uum.edu.my/9588/1/sh.pdf; http://repo.uum.edu.my/9588/

Data mining is a popular technique to find hidden in formation in database. A lot of studies have been conducted to show that data mining technique is a powerful tool that is capable to provide highly targeted information to support decision making and forecasting for business, science, health, education, industry and others.In education sector, data mining techniques is normally used to predict student’s performance in certain courses, to forecast the lecturers performance at the university and others. Indirectly, these techniques contribute towards a better quality education management, as well as to assist the education institution managing the administrative tasks effectively. This study aims to develop prediction models for determining the program undertaken at Matriculation College based on the student’s background, academic achievement as well as personality traits.To accomplish this, NN model known as multilayer perceptron with back propagation learning and regression model were employed. The findings show that Neural Network has more accuracy percentage than Logistic Regression.It also presents the existing relationship between students‘ achievements, personality traits and course undertaken.