fish_class

date: Sat, 13 Oct 2007 18:07:27
SVN revision: 146

units used: 512
train objects: 5

data properties:
#bins
phi_y 36
y 36
x 36
z 36
id 10

parameters:
n_bins 36
reg_train_fraction 0.5
phase_b_normalised True
KNN_k 8
class_train_fraction 0.5
class_test_fraction 0.03

Histograms


hd png

hd png

hd png

hd png

hd png

Classification

Classifier on old Objects


hd png

hd png

KNN (k=8) Classifier

744 samples, 17 errors
97.72 % hit rate, 2.28 % error rate

classification Matrix:
obj. 1 obj. 2 obj. 3 obj. 4 obj. 5
classified as obj. 1 150 0 0 0 0
classified as obj. 2 0 151 3 1 0
classified as obj. 3 1 1 148 4 2
classified as obj. 4 1 1 2 138 1
classified as obj. 5 0 0 0 0 140
LaTeX

Mahalanobis Classifier

744 samples, 40 errors
94.62 % hit rate, 5.38 % error rate

classification Matrix:
obj. 1 obj. 2 obj. 3 obj. 4 obj. 5
classified as obj. 1 152 1 0 1 0
classified as obj. 2 0 152 12 20 0
classified as obj. 3 0 0 141 6 0
classified as obj. 4 0 0 0 116 0
classified as obj. 5 0 0 0 0 143
LaTeX

Classifier on new Objects


hd png

hd png

KNN (k=8) Classifier

749 samples, 6 errors
99.20 % hit rate, 0.80 % error rate

classification Matrix:
obj. 6 obj. 7 obj. 8 obj. 9 obj. 10
classified as obj. 6 147 0 0 0 0
classified as obj. 7 0 145 0 1 1
classified as obj. 8 1 0 151 0 0
classified as obj. 9 0 0 0 152 1
classified as obj. 10 2 0 0 0 148
LaTeX

Mahalanobis Classifier

749 samples, 89 errors
88.12 % hit rate, 11.88 % error rate

classification Matrix:
obj. 6 obj. 7 obj. 8 obj. 9 obj. 10
classified as obj. 6 150 30 22 4 26
classified as obj. 7 0 109 0 0 0
classified as obj. 8 0 6 129 1 0
classified as obj. 9 0 0 0 148 0
classified as obj. 10 0 0 0 0 124
LaTeX

Classifier on all Objects


hd png

hd png

KNN (k=8) Classifier

1493 samples, 49 errors
96.72 % hit rate, 3.28 % error rate

classification Matrix:
obj. 1 obj. 2 obj. 3 obj. 4 obj. 5 obj. 6 obj. 7 obj. 8 obj. 9 obj. 10
classified as obj. 1 145 0 0 0 0 6 0 0 0 1
classified as obj. 2 0 147 3 1 0 0 0 0 0 0
classified as obj. 3 1 2 146 3 1 0 3 4 0 2
classified as obj. 4 1 2 3 138 0 0 1 0 0 0
classified as obj. 5 0 0 0 0 140 0 1 0 0 0
classified as obj. 6 3 0 0 0 0 144 0 0 0 0
classified as obj. 7 0 0 0 1 2 0 140 0 1 1
classified as obj. 8 0 2 0 0 0 0 0 147 0 0
classified as obj. 9 1 0 0 0 0 0 0 0 152 1
classified as obj. 10 1 0 1 0 0 0 0 0 0 145
LaTeX

Mahalanobis Classifier

1493 samples, 167 errors
88.81 % hit rate, 11.19 % error rate

classification Matrix:
obj. 1 obj. 2 obj. 3 obj. 4 obj. 5 obj. 6 obj. 7 obj. 8 obj. 9 obj. 10
classified as obj. 1 130 0 0 0 0 0 1 0 0 0
classified as obj. 2 0 146 12 19 0 0 0 0 0 0
classified as obj. 3 0 0 136 6 0 0 0 0 0 0
classified as obj. 4 0 0 0 114 0 0 0 0 0 0
classified as obj. 5 0 0 0 0 141 0 1 0 0 0
classified as obj. 6 21 5 4 2 0 150 30 22 4 26
classified as obj. 7 0 0 0 2 1 0 108 0 0 0
classified as obj. 8 1 2 0 0 1 0 5 129 1 0
classified as obj. 9 0 0 0 0 0 0 0 0 148 0
classified as obj. 10 0 0 1 0 0 0 0 0 0 124
LaTeX