Page 327 -

Index 317

linkage transformation, 37
average between groups, 62 out-degree, 248
average within groups. 62 outlier, 42
complete, 58, 62 output neuron, 155
single, 60
Ward, 63 PAC, 200
LMS, 149 parameter estimation, 97
local minimum, 157, 177 parsing tree, 252
Logistic distribution, 100 partition method, 128
Parzen window, 110
Mahalanobis pattern
classifier, 85 acquisition, 16
distance, 36, 99 classification, 4
majority vote, 233 description, 9
margin of separation, 2 15 primitive, 243
match graph, 280 recognition, 1
matching score, 269, 276 regression, 7
max vote, 233 patterns, 12
maximum likelihood, 98 PDL, 253
merit perceptron, 159
criterion, 12 1 rule, 160
index, 7 1 piecewise linear, 244
metric, 29 pooled covariance, 103
MLP, 167 post-processing, 17, 176
modular network, 232 predictive modelling, 235
momentum factor, 175 pre-processing, 17, 176
multidimensional scaling, 67 prevalence, 90
mutation, 209 prevalence threshold, 91
primitive, 243
nearest neighbour, 1 14 principal component, 39, 65
network probabilistic
complexity, 169 neural network, 113,212
pruning, 191 relaxation, 27 1
weight, 149 production rule, 251
node prototype state, 230
splitting, 138 pseudo
node impurity, 136 dimension, 198
node set, 248 inverse, 149,207
noise adding, 178 shattering, 197
norm
Chebychev, 29 quadratic
city-block, 29 classifier, 24, 127
Euclidian, 29 decision function, 23
Mahalanobis, 32 form, 32
Minkowsky, 30
power, 29 radial
null string, 247 basis functions, 168, 2 12, 222
layer, 2 13
optimal hyperplane, 2 16 ranking index, 185
optimum classifier, 99 RBF, 168,213
orthonormal
recurrent network, 169
matrix, 38 regular grammar, 256