作者: J. Miguel Marín , Genaro Sucarrat
DOI: 10.1007/978-88-470-2342-0_33
关键词:
摘要: We study the properties of two methods for financial density selection Skewed Exponential Power (SEP) distribution. The simulations suggest can be great use in practice, since recovery probabilities are sufficiently high finite samples. For first method, which simply consists selecting a by means an information criterion, Schwarz criterion stands out it performs well across categories, and particular when Data Generating Process (DGP) is normal. In smaller samples that our second General-to-Specific (GETS) selection, improve rate predictable ways changing significance level. This useful because enables us to increase chosen category, if one wishes do so.