The Development of a Prediction System for Student Learning Progress Based on Artificial Neural Networks (A Case Study on Study Program of Mathematics and Statistics – The Faculty of Mathematics and Natural Sciences, Universitas Terbuka)

When new students enroll at the university, they need to fill application forms that incorporate any information about themselves such as academic background, permanent mailing address, gender, date of birth, occupation, marital status, etc. However, this information is not utilized well enough by the university to help in overcoming low graduation rates. This research applies Artificial Neural Network (ANN) Multilayer Perceptron to predict progress learning of students using several parameters such as individual parameter (age, gender), environment parameter (marital status, occupation), and academic parameter (entry semester --odd or even--, grade point average in the first semester, the number of credit hours in the first semester, a cumulative grade point average, the total number of semesters completed at the university, and the study program) at Universitas Terbuka (The Indonesian Open University). An Artificial Neural Network (ANN) is an information processing paradigm that is inspired by the way biological nervous systems, such as the brain, process information. The key element of this paradigm is the novel structure of the information processing system. It is composed of a large number of highly interconnected processing elements (neurons) working in unison to solve specific problems. ANNs, like people, learn by example. An ANN is configured for a specific application, such as pattern recognition or data classification, through a learning process. Learning in biological systems involves adjustments to the synaptic connections that exist between the neurons. This research also applies sensitivity analysis method to measure the influence of individual input parameter on any one of possible outcomes. The experiment had been conducted using student data collection from two study program (Mathematics and Statistic). The data was collected from 3517 students, with 81 students finished their degrees. The experiments used 50% of data as training set, 25% as validation set and 25% testing set. Our experiments and simulation results indicated that the sensitivity analysis method was a potential tool to reduce the complexity of ANN Multilayer Perceptron and to increase the generalization. Generalization is the recognition level of neural network toward the given pattern. The results showed that the generalization of the prototype had an accuracy of 0.992 in predicting the correct outcomes of student graduation