Prediction of Bond Planning Based on Customer Characteristics in Indonesia Using the Random Forest Algorithm

Abstract

This research aims to identify industries receptive to wealth management products like bond insurance and to pinpoint the key factors influencing bond insurance premium closures. Given Indonesia's low life insurance penetration and density compared to other Asian countries, the study uses a purposive sample of 199 data points from 38 companies across various industries. This research found that the most influential factors in bond purchases are domicile, the second car, industry type, job position , housing type . The methodology that is used in thisresearch is called CRIPS-DM (Cross Industrial Standards Program Data Mining). The first steps what is the purpose of the organization, and the second is what data that needed, and then continue to data preparation, after that modeling, after modeling, it will make an interpretation of the result, and the final steps is deployment, it will plan how it will be implemented in the real world, and the accuracy score from this model is 98 %. From the result of the projection closing bond insurance from each industry, it can be concluded that the most industry that closed the bond insurance is Banking Industry, the second is from insurance and the third is education and the next is education, retail, bond, manufacturing and finance, hospitality, legal, publishing, technology and government and service industries.