Credit risk assessment is of paramount importance in the financial industry, guiding lenders in making informed decisions regarding loan approvals and interest rates. This research presents a holistic analysis of a dataset, encompassing the application of logistic regression to construct a probability of default (PD) model, estimating loss given default (LGD), and introducing the innovative use of Weight of Evidence (WoE) and Information Value (IV) to assess and rank the predictive power of diverse risk factors.
Additionally, a novel quantization technique is proposed to transform continuous variables into discrete counterparts. Upon concluding this paper, readers are equipped to extend their research efforts by applying the quantization technique to all available risk factors within the dataset, subsequently comparing the predictive performance against their own models.