MACHINE LEARNING PREDICTIVITY APPLIED TO INDONESIAN SMARTPHONE USERS’ CREDITWORTHINESS AFTER COVID-19 PANDEMIC
In this research work, we used machine learning techniques to predict the creditworthiness of smartphone users in Indonesia after the COVID-19 pandemic. Principal Component Analysis (PCA) and K-means algorithms were used for dimensional reduction and clustering using a dataset of 803 respondents consisting of twelve questions to smartphone users in Indonesia after the COVID-19 pandemic. To classify the creditworthiness of smartphone users in Indonesia, the four different classification algorithms (Logistic Regression, Support Vector Machine, Decision Tree, and Naïve Bayes) were tested. The tests carried out included testing for accuracy, precision, recall, F1-score, and Area Under Curve Receiver Operating Characteristics (AUCROC) assessment. When compared to other models, the Logistic Regression algorithm outperforms them. The findings of this study also provide new information about the most influential and non-influential variables based on the twelve questions posed to smartphone users in Indonesia, which can assist financial institutions, particularly banks, in assessing the creditworthiness of prospective customers after COVID-19 pandemic.
Creditworthiness, Smartphone, Machine Learning, COVID-19 Pandemic