Credit rating analysis with support vector machines and neural networks: a market comparative study PDF
Book Description:
Abstract Corporate credit rating rating evaluation has attracted lots associated with research interests inside literature. Recent studies have shown that will Artificial Intelligence (AI) methods achieved better performance when compared with traditional statistical methods. This specific article introduces any relatively new machine learning technique, support vector machines (SVM), in order to the problem in attempt in order to provide any model along with better explanatory power. We used backpropagation neural network (BNN) since a benchmark along with obtained prediction accuracy around 80% pertaining to both BNN along with SVM methods pertaining to the United States and Taiwan markets. On the other hand, only slight improvement of SVM was observed. Another direction with the research is to improve the interpretability from the AI-dependent models. We applied recent research final results in neural network style interpretation and also obtained relative importance of the input monetary variables coming from the neural network versions. Based about these final results, we conducted any market comparative investigation on the particular differences regarding determining factors in the United States along with Taiwan markets.
Credit rating analysis with support vector machines and neural networks: a market comparative study PDF:
http://www.personal.psu.edu/faculty/h/u/huz2/Zan/papers/credit.dss.pdfQuick View PDF Download PDF
