Penerapan Metode Klasifikasi Multinomial Naive Bayes

(Studi Kasus: PT Prudential Life Samarinda Tahun 2019)

  • Rival Rinaldi Laboratorium Statistika Komputasi FMIPA Universitas Mulawarman
  • Rito Goejantoro Laboratorium Statistika Komputasi FMIPA Universitas Mulawarman
  • Syaripuddin Syaripuddin Laboratorium Matematika Komputasi, FMIPA Universitas Mulawarman

Abstract

Life insurance is a risk management service provide payment to policyholders in the event of a disaster that has been stipulated in the agreement. A classification system needs to be done to facilitate the company in making decisions to provide policies to customers. One system that can be used is multinomial Naive Bayes. Multinomial Naive Bayes is a simple probabilistic classification that has more than two groups or categories. An algorithm using Bayes theorem assumes all independent variables. The aim of this study is to obtain an accuracy level of 5 different proportions with the Naive Bayes multinomial method used in insurance customer payment status data. The data used is the customer data of PT. Prudential Life Samarinda in 2019 with the status of current premium payment, substandard and non-current and using 5 independent variables, namely income, age, amount of premium payment, sex and employment. The results of the measurement of classification accuracy using APER status premium payment on insurance customer data of PT. Prudential Life 2019 Naive Bayes multinomial method showed 22,96% misclassification at 50:50 proportion, at the proportion of 60:40 there were 21,43% misclassification, at the proportion of 70:30 there were 19,05% misclassified, at proportions 80:20 had a misclassification of 14,29%, and a proportion of 90:10 has a misclassification of 7,14%.

Downloads

Download data is not yet available.
Published
2021-12-30
How to Cite
RINALDI, Rival; GOEJANTORO, Rito; SYARIPUDDIN, Syaripuddin. Penerapan Metode Klasifikasi Multinomial Naive Bayes. EKSPONENSIAL, [S.l.], v. 12, n. 2, p. 111-118, dec. 2021. ISSN 2798-3455. Available at: <https://jurnal.fmipa.unmul.ac.id/index.php/exponensial/article/view/803>. Date accessed: 10 may 2024. doi: https://doi.org/10.30872/eksponensial.v12i2.803.
Section
Articles