PENERAPAN ALGORITMA NAÏVE BAYES UNTUK PENENTUAN RESIKO KREDIT BANK

Abstract

This study aims to classify in determining the value of bank credit risk. The data is obtained from a data bank with a population of 900 data where this data is divided into 800 training data and 100 testing data. The attributes used are 5 namely, annual income, Active KPR, Loan Duration, Number of Dependents and Average Overdue. The method used in this research is the nave Bayes algorithm which is processed using R Studio software. For manual calculation of training data used 100 data from 800 training data, and testing data taken 5 data from 100 testing data. Meanwhile, the data processed by R Studio has an accuracy of 0.997 or 99%. Sensitivity class 1 100%, Sensitivity class 2 98%, Sensitivity class 3 100%, Sensitivity class 4 100% and Sensitivity class 5 100%. Specificity class 1 100%, Specificity class 2 100%, Specificity class 3 100%, Specificity class 4 100% and Specificity class 5 99%.
Keywords: Naïve Bayes, Credit Risk
ABSTRAK
Penelitian ini bertujuan untuk melakukan klasifikasi dalam menentukan nilai resiko kredit bank. Data didapat dari data bank dengan populasi 900 data dimana data ini dibagi menjadi 800 data training dan 100 data testing. Atribut yang digunakan sebanyak 5 yaitu, pendapatan setahun, KPR Aktif, Durasi Pinjaman, Jumlah Tanggungan dan Rata-rata Overdue. Metode yang digunakan dalam penelitian ini adalah algoritma naïve bayes yang di proses menggunakan software R Studio. Untuk perhitungan manual data training yang digunakan 100 data dari 800 data training, dan data testing diambil 5 data dari 100 data testing. Sedangkan data yang di proses dengan R Studio memiliki accuracy 0.997 atau 99%. Sensitivity class 1 100%, Sensitivity class 2 98%, Sensitivity class 3 100%, Sensitivity class 4 100% dan Sensitivity class 5 100%. Specificity class 1 100%, Specificity class 2 100%, Specificity class 3 100%, Specificity class 4 100% dan Specificity class 5 99%.
Kata Kunci: Naïve Bayes, Resiko Kredit