Research Article

Ensemble Learning Based Multiple Kernel Principal Component Analysis for Dimensionality Reduction and Classification of Hyperspectral Imagery

Table 4

Overall accuracy, OA (%), and Kappa statistic, , of the KSC data set with different models. The best scores for each class are highlighted in bold face font, and the second best is underlined.

FeatureRawPCAKPCARBFKPCALapKPCACauKPCAHISTA-KPCAM-KPCA

# of features176385618551412

SVMC534.66279.17430.54663.98181.02346.69117.38165.69
paramsσ3.779.8734.95 5.5810.695.585.615.01

197.5092.6497.5095.6696.5896.3298.0398.29
294.2484.7793.8390.1291.7791.3694.2495.47
393.7589.8493.3691.8093.7593.7593.3697.26
478.1749.2180.1671.4378.1775.4081.3593.25
577.6456.5260.8755.2861.4957.7678.8888.82
675.1150.6658.0855.0275.1158.9582.5386.90
793.3473.3389.5291.4390.4891.4394.2995.24
896.2978.6594.2090.2695.1391.1896.7596.28
999.0497.5098.6596.7398.6597.5099.2399.61
1010091.5899.7597.7799.2697.7710099.26
1199.0597.8599.0598.0999.0598.8199.0599.28
1299.0681.1196.8295.2397.4295.6399.0199.20
1310099.8910099.8999.8999.8999.89100

OA95.5186.5393.7891.6594.3492.5996.1497.52
0.95220.84930.93470.91060.94070.92130.96010.9724