Comparison between classifiers - Eimeria of domestic fowl

The confusion matrices of both methods are presented below:

  • Parametric classifier (similarity scores)

    ACE
    MAX
    BRU
    MIT
    PRA
    TEN
    NEC
    ACE
    91.80
    0.00
    0.00
    1.34
    0.00
    0.80
    5.88
    MAX
    0.00
    97.09
    2.91
    0.00
    0.00
    0.00
    0.00
    BRU
    0.00
    0.48
    96.41
    0.00
    0.48
    1.91
    0.72
    MIT
    1.79
    0.00
    0.00
    85.08
    9.25
    2.09
    1.79
    PRA
    0.00
    0.00
    2.65
    8.75
    74.01
    6.63
    7.96
    TEN
    0.00
    0.00
    1.29
    0.97
    5.79
    87.14
    4.82
    NEC
    6.56
    0.00
    1.16
    2.70
    13.90
    7.34
    68.34

    Overall correct species assignment: 85.72%

  • Non-parametric classifier (probability scores)

    ACE MAX BRU MIT PRA TEN NEC
    ACE 87.70 0.00 0.00 1.34 0.54 0.80 9.63
    MAX 0.00 96.12 3.88 0.00 0.00 0.00 0.00
    BRU 0.00 0.72 94.98 0.00 0.24 3.11 0.96
    MIT 1.49 0.00 0.00 86.27 9.25 0.30 2.69
    PRA 0.00 0.00 1.34 11.67 64.46 15.39 7.16
    TEN 1.61 0.00 1.93 1.29 7.72 76.53 10.93
    NEC 5.79 0.00 1.93 5.79 12.74 18.15 55.60

    Overall correct species assignment: 80.24% 

-
   © 2007 - A. Gruber, C.A.B. Pereira, L.F. Costa, & C.A.B. Castañón