1. NEURAL NETWORKS FOR CREDIT RISK MANAGEMENT: A CASE STUDY IN THE CAR FINANCING INDUSTRY
Authors: ANNA MATUSZYK , HSIN-VONN SEOW , MARTIN MULLER , STEFAN LESSMANN
Abstract: The paper aims at examining the degree to which artificial neural networks are a suitable approach to aid risk management in the car financing industry. More specifically, we empirically compare a classic feedforward neural network to a recently proposed extreme learning machine. To that end, we employ a real-world credit data set from a leading car financing company in Poland is used to assess each classifier’s accuracy. To systematically study the suitability of the two methods, our study comprises multiple experimental factors including the type of the neural network, whether or not it is embedded into an ensemble learning framework, and strategies to mitigate class imbalance.
Keywords: credit risk, decision support, data analytics