|
A decision tree, which was trained using B/S and P/L of
companies that have been defaulted (D) and have not been defaulted
(N), by using SONAR.
A regression tree, which was trained using the New York markets
data, by using SONAR. The data contains numeric attributes such as
"BPS" (British pound sterling), "GDM" (Deutschmark), "YEN" (Japanese
yen), "TB3M" (3-month Treasury bill yields), "TB30Y" (30-years
Treasury bill yields), "GOLD" (gold price per ounce) and "SP500"
(Standard and Poors index). The tree is constructed so that
variance of the SP500 is minimized.
|
