| Summary: | 碩士 === 國立彰化師範大學 === 企業管理學系 === 95 === The objective of this study is in developing a reliable credit rating forecasting model for lenders as the important criteria of enterprise risk assessment to reduce the loss resulting from financial distress. We adopt 453 High-Tech companies from TEJ as samples and use 35 financial indices as our input data. This study adopts three measurements, including Support vector machines, logistic model and back-propagation neural networks, to analyze credit rating forecasting; furthermore, we combine them with Genetic Algorithm so as to compare the six different results. Empirical results suggest that SVM-based model with Genetic Algorithm outperforms other models. The model has 85.3261% accuracy rate.
|
|---|
