PEMBELAJARAN MESIN UNTUK MENILAI KELAYAKAN KREDIT PROYEK RETROFIT: MULTINOMIAL LOGIT

Abstract

Creditworthiness assessment was one of the first areas to apply machine learning techniques in economics. The creditworthiness of retrofit protection was vital for ESCO in determining the credit scoring. This study aimed to develop a retrofitting assessment model to utilize machine learning with multinomial logistic (MNL) and life cycle cost analysis (LCCA). This study aims to provide an evaluation of creditworthiness models from the evaluation of financing alternative in Indonesia's energy efficiency industry. The goal was to reduce the total of prediction error, which comprised bias, variance, and fundamental error. The findings demonstrated that machine learning approaches might yield significantly greater prediction accuracy. In addition, machine learning is also expected to automatically capture the nonlinear relationship between input features and selected results. This study is also expected to draw on ideas from machine learning to develop an enhanced model for retrofitting creditworthiness research and suggest new research directions.