Table 5.

Permutation-based validation of the classification signal using PACE methodsa

ParameterIMAC (%)CM10 (%)
BKVAN versus ARBKVAN versus ControlAR versus ControlBKVAN versus ARBKVAN versus ControlAR versus Control
SVM + all peaks
    ACE29.5623.1439.1441.6135.1640.56
    MACE48.6147.9650.2548.7748.2750.34
    95%45.1745.1147.3245.9445.3347.59
    99%44.4444.4746.4545.2245.0547.18
SVM + 100 t test
    ACE16.8311.3317.4525.9430.9830.37
    MACE48.5748.0650.2548.8048.4250.31
    95%46.7246.3348.5046.8946.2548.10
    99%46.1145.5947.4545.6745.6047.41
RF
    ACE27.5016.7622.9135.2834.6237.59
    MACE45.2342.8850.7045.5844.1051.01
    95%43.6741.3348.8643.9442.6649.12
    99%42.6740.5347.9542.8941.4748.80
  • a The statistics shown include ACE and mean classification error (MACE) for the error distribution under the random permutation of labels and its 95th and 99th percentiles of this distribution.