Research on Credit Risk Quantification and Optimization Strategies for Micro, Small and Medium Enterprises (MSMEs)

The swift growth in China's finance market has highlighted the important role of micro, small, and medium-sized enterprises (MSMEs) in contributing to job creation, improving livelihood, and generating stable government income. However, MSMEs have always encountered a series of serious financing difficulties due to their small size, low credit ratings, and unavailability of collateral. These difficulties not only hinder their growth but also complicate credit resource management and credit risk assessment for banks. Thus, building an effective credit assessment system is a crucial measure to alleviate financing difficulties for MSMEs. This paper recommends building a credit risk assessment and strategy optimization model for MSMEs specifically, with a decision tree approach and K-means cluster algorithms. Initially, the research identifies influential variables affecting credit risk and constructs a quantitative model according to enterprise transaction information, finance factors, and credit ratings. Subsequently, using K-means cluster, firms are classified into 3 levels of credit risk (high, medium, and low) to enable diversified credit strategies to be recommended. Results show that the model reduces loan defaults to a great extent and increases credit resource allocation efficiency. Additionally, using non-financial information strengthens the validity of the model, with empirical results displaying high accuracy and applicability. With optimization, not only is the default rate for banks reduced, but loyalty for low-risk firms is increased, making overall credit decision-making more rational. Subsequent research can draw on a range of sources of data, from instantaneous transaction history to external market information and business social network interactions, and use the K-means++ algorithm to improve cluster centers initialization and thus improve model accuracy and stability. Additionally, use of real-time data enables ongoing observation and measurement of corporate credit conditions, which can in turn result in further credit strategy refinement and improve the ability of financial institutions to manage risks.