Prediksi Tingkat Keberhasilan Studi Kinerja Santri Menggunakan Algoritma C 5.0

Abstract

The success of pesantren education institutions can be measured by the success of their students. By predicting the possible outcomes of the learning process based on prediction results can help an Islamic boarding school, by adjusting the factors that contribute and influence the success rate of students' performance studies. And by utilizing data mining techniques that can be used to increase the level of success and reduce the failure of students. this can greatly help pesantren educational institutions to improve their graduates 'skills, because data mining is the best solution to find hidden patterns and can predict the success of students' performance studies. This research presents a model based on decision tree classification algorithm C 5.0 used in this model with alumni tracer study filled by santri alumni. In this study also used the k-folds cross validation test scenario with k values of 2,3,6,10 and 15 with a total of 300 alumni data and 84 data used for validation tests without cross validation. Determination of the criteria for the classification results using a confusion matrix form the measurement of the classification results obtained, namely the highest value in this study is 95% resulting from 15 folds the scenario 1. And form the results of testing the validation data without cross validation, the corresponding results are 73.81%, when compared to the k-folds, there was an increase of 21.19% and it can be ignored that the C 5.0 algorithm is able to classify well. So that pesantren educational institutional can provide a foundation in the arrangement for their students in deciding the right school choice.