Research Article

Weighted Decoding for the Competence Reliability Problem of ECOC Multiclass Classification

Table 4

Classification accuracy of DRCW-OVO and WCR.

DatasetSVMTreec
DRCWWCRDRCWWCR

Ecoli66.67 ± 2.9369.01±2.7372.91±0.8669.64 ± 1.15
Glass97.26±3.7393.47 ± 3.7290.54 ± 7.0892.98±2.93
Satimage85.84 ± 0.5486.14±3.8383.93 ± 1.4784.01±0.79
Segment93.25 ± 0.9094.87±1.1493.94 ± 1.7194.68±1.35
Vehicle79.31 ± 3.3879.54±2.8765.97 ± 2.6267.50±1.84
Vowel98.69±0.8485.66 ± 3.2782.73±2.9282.02 ± 5.41
Yeast57.34 ± 2.3258.48±2.6051.15 ± 3.8851.87±3.02
Zoo81.02 ± 4.1283.26±3.8971.23 ± 13.9174.58±3.26