Prediction of General High School Exam Result Level Using Multilayer Perceptron Neural Networks

摘要

The general high school exam is considered one of the most important exams for the student. The achievement of this academic qualification enables him to build his future career and determine the course of his life through joining a bachelor program in a university based on the marks obtained in this exam. On another hand, the universities determine the admission rates for each discipline and program depending on the general high school exam results in each year. Some universities with some specialization start the acceptance from high-level grade of high school exam and decrease it gradually to reduce the acceptance grade level depending on the demand for a specific specialization. A smart prediction model of the general high school exam results in next year will help the universities to determine the level from the beginning of the acceptance process. In this paper, we present a prediction model which has the ability to analyze previous patterns of general high school exam result and use them to predict the future results. The time series dataset consists of general high school exam results for the previous eleven years in Palestine, which was used as a target for the Multilayer Perceptron Neural Networks with Backpropagation learning algorithm (MLPBP) to predict the future general high school exam results. The accuracy of the prediction results of the proposed model is significant as it appears in the result of Minimum Mean Square Error (MSE). Moreover, the prediction result of each level in the next year is very accurate regarding the compiled pattern of the historical data.