Abstract:
"This study focuses on the credit risk rating prediction using the XGBoost and GMM for the finance industry. The highlighted project from the Department of Marketing and Relationship Management aims at assessing customer information for better credit risk ratings and better categorization of bulk customers. Specifically, for this paper’s purpose, this research envisions to synthesize the strengths of XGBoost in advanced machine learning and the capacity of GMM for clustering to optimize decision making in the financial institution through assessing accuracy in modeling customers’ behaviour and financial risks.
However, even with all these developments in credit risk management a lot of traditional methods fail in terms of properly assessing the customer records and performing accurate segmentation. These limitations are handled in this research utilizing XGBoost and GMM since they have revealed effectiveness in other researches although they have not been implemented in this particular field of application. The use of these techniques is realized with the intention of filling this gap with a more flexible and elaborate credit risk assessment that enlightens the financial institutions in handling of risks related to credit.
From the results obtained from the benchmark tests, the integrated model of XGBoost and GMM has enhanced better performance in credit risk prediction. The findings display the model successfully achieving a 96% percent accuracy level after training, which underlines the model’s strength in assessing customer credit risk. This high level of accuracy has a positive implication of ensuring accurate credit risk based on different customers’ history and their reliability in enhancing financial decisions. The desired enhancement of the model accuracy from the test sample to the trained model explains why incorporating XGBoost with GMM can support credit risk analysis.
"
