A big data analytics method for assessing creditworthiness of SMEs: Fuzzy equifinality relationships analysis

Nowadays, many financial institutions are beginning to use Big Data Analytics (BDA) to help them make better credit underwriting decisions, especially for small and medium-sized enterprises (SMEs) with limited financial histories and other information. The various complexities and the equifinality problem of Big Data make it difficult to apply traditionalstatistical techniques to creditworthiness evaluation, or credit scoring. In this study, we extend the existing research in the field of creditworthiness assessment and propose a novel approach based on neighborhood rough sets (NRSs), to evaluate and investigate the complexities and fuzzy equifinality relationships in the presence of Big Data. We utilize a real SME loan dataset from a Chinese commercial bank to generate interval number rules that provide insight into the fuzzy equifinality relationships between borrowers’ demographic information, company financial ratios, loan characteristics, other non-financial information, local macroeconomic indicators and rated creditworthiness level. In addition, the interval number rules are used to predict creditworthiness levels based on test data and the accuracy of the prediction is found to be 75.44%. One of the major advantages of using the proposed BDA approach is that it helps us to reduce complexity and identify equivalence relationships when using Big Data to assess the creditworthiness of SMEs. This study also provides important implications for practices in financial institutions and SMEs.