Analysis of Credit Default Prediction Based on Logistic Regression, Random Forest and KNN Model

Addressing the prevalent issue of limited access to loans for individuals with inadequate or non-existent credit histories is imperative to foster financial inclusion and safeguard vulnerable populations from unscrupulous lenders. With the help of unique characteristics, this research seeks to improve financial accessibility for the unbanked population by creating a safe and satisfying financing environment. The paper uses an imbalanced dataset that was gathered from Kaggle to examine how machine learning techniques can be applied to forecast credit risk. specifically, Logistic Regression, Random Forest, KNN, and Synthetic Minority Over-sampling Technique (SMOTE) are utilized in this study. The findings reveal an enhancement in accuracy following the resolution of the imbalanced dataset challenge. By employing these machine learning models, the research seeks to not only bridge the credit gap for the underserved but also mitigate the risks associated with lending to individuals lacking conventional credit histories. A thorough strategy for managing credit risk in the context of financial inclusion is shown by the investigation of different algorithms and the application of methods. These results contribute to the discourse on leveraging machine learning to create more equitable lending practices, ensuring that the unbanked population can access financial resources responsibly.