| Summary: | 碩士 === 國立東華大學 === 應用數學系 === 98 === This paper used the Gaussian mixture model to find credit risk. The author referred to Fraley and Raftery (2002), used the covariance that parameterized by eigenvalue decomposition and got ten models. As for the variables, the author extracted 22 variables from Altman (1968), Shumway (2001), Duffie (2007), Compbell (2008), and several financial related books. The author selected five variables and collocated with the ten Gaussian mixture models. The result indicated that the VEI model performed well combined with the variables that the author found. Compared with the classification of TEJ TCRI, the empirical result indicated that the author’s classification result had better classified result.
|
|---|
