IMPLEMENTASI METODE NAIVE BAYES UNTUK KLASIFIKASI GOLONGAN PEMBAYARAN SYAHRIYAH DI PONDOK PESANTREN DARUSSAADAH

Currently, Pondok Pesantren Darussa’adah continues to utilize a traditional approach in managing
its operational systems, particularly in the administration of syahriyah (monthly tuition) payments.
The payment system at this pesantren implements a cross-subsidy mechanism, where students from
more economically advantaged backgrounds help subsidize those from less privileged families.
However, the classification process for syahriyah payment tiers remains manual and is based on a
form that evaluates Monthly Income, Number of Dependents, and Outstanding Loans. This form is
submitted alongside the new student registration form. Such a conventional method is considered
inefficient and prone to human error in the classification process.
Therefore, this study aims to classify syahriyah payment categories using the Naive Bayes method.
A dataset comprising 1,000 student records from Pondok Pesantren Darussa’adah was utilized, with
70% allocated for training and 30% for testing. The study evaluates and compares the classification
performance of four different models employing the same Naive Bayes technique. The best
performing model achieved an overall accuracy of 96%, with precision and recall also at 96%.
Class-wise accuracy results indicate 100% accuracy for the “Capable” category, 96% for the
“Middle” category, and 93% for the “Underprivileged” category.