Research Article

Combination of Biorthogonal Wavelet Hybrid Kernel OCSVM with Feature Weighted Approach Based on EVA and GRA in Financial Distress Prediction

Table 5

Results of FWOCSVM and OCSVM based on different kernels.

Indicator setKernel functionFWOCSVMOCSVM
CCRCCR

Linear79.2%71.5%19.2%77.3%70.4%20.2%
Poly78.3%71.5%20.5%76.7%68.7%21.5%
Sigmoid79.3%72.0%20.2%77.5%70.8%21.0%
RBF80.8%76.3%18.5%77.8%71.5%20.2%
Morlet80.8%76.8%18.4%78.8%73.3%19.6%
Coif381.7%77.2%17.6%80.5%74.3%18.8%
Cdf9/782.5%77.3%16.4%80.8%75.2%18.0%
87.5%81.2%11.2%84.3%79.5%15.9%

Linear80.0%74.1%18.8%78.2%72.5%19.9%
Poly79.0%73.6%18.7%77.5%71.0%21.1%
Sigmoid80.2%73.1%18.6%78.3%72.5%20.4%
RBF82.7%78.7%16.7%79.2%74.4%19.8%
Morlet83.1%79.7%16.0%80.0%76.3%19.2%
Coif383.3%81.2%16.2%81.7%77.3%17.4%
Cdf9/784.2%80.7%15.2%82.3%77.3%16.7%
90.0%91.4%10.3%85.8%79.7%12.9%

Linear78.3%70.1%19.7%76.5%69.6%22.1%
Poly77.5%69.1%22.1%75.8%67.6%22.5%
Sigmoid77.7%69.7%21.4%76.7%70.1%22.0%
RBF80.2%73.0%19.7%77.5%71.1%21.2%
Morlet79.5%74.9%19.1%78.3%72.1%20.4%
Coif380.8%75.8%17.9%79.2%73.0%19.6%
Cdf9/781.7%77.3%17.2%80.0%73.5%18.7%
85.8%83.9%13.6%83.3%78.9%15.8%