Index
Neural Smithing: Supervised Learning in Feedforward Artificial Neural Networks
Russell D. Reed and Robert J. Marks II
Copyright © 1999 Massachusetts Institute of Technology
Index
C
Capacity,
41
-
47
.
See also
Representational capabilities
Cascade-correlation algorithm,
201
-
204
,
268
Cauchy's method,
150
,
165
-
166
Cell bodies,
1
CGD (conjugate gradient descent),
166
-
169
,
179
,
182
,
252
Chain rule,
53
,
66
-
67
Classical optimization methods.
See also
Genetic algorithm
;
Optimization methods
best-step steepest descent,
165
-
166
conjugate gradient descent,
166
-
169
,
179
,
182
,
252
deterministic,
158
,
159
-
175
Gauss-Newton,
130
,
172
gradient-based,
158
,
163
-
175
,
182
Hooke-Jeeves pattern search,
160
-
161
Levenberg-Marquardt,
128
,
144
,
172
-
173
,
182
Nelder-Mead simplex search,
161
Newton's,
128
,
145
,
152
,
170
-
172
,
179
,
294
-
295
Powell's conjugate direction,
161
-
163
quasi-Newton,
174
-
175
simulated annealing,
157
,
176
-
178
Classical optimization methods (
cont.
)
stochastic,
158
,
175
-
179
Classification
initialization and,
107
-
109
,
110
regression versus,
156
-
157
Clipped linear function,
317
Cluster centers,
107
Clustering,
158
Competition, and pruning,
232
-
233
Complexity,
241
-
242
,
244
-
247
Computation speed,
4
-
5
Conjugate gradient descent (CGD),
166
-
169
,
179
,
182
,
252
Conjugate methods,
144
,
152
,
161
-
163
Connected, fully,
3
Constraints,
240
-
241
,
244
-
245
Constructive methods,
197
-
199
,
217
cascade-correlation,
201
-
204
,
268
dynamic node creation,
199
-
201
generalization and,
268
Marchand's,
209
-
212
meiosis networks and,
212
-
213
node splitting,
212
-
215
pruning and,
198
,
201
tiling,
206
-
209
upstart,
204
-
206
Voronoi tessellation,
215
-
216
Convergence time.
See
Training time
Convex regions,
33
Convolution,
277
-
281
,
311
-
314
Cost function,
7
,
155
-
156
.
See also
Error function
;
Objective function
Cross-entropy error function,
9
,
50
,
110
,
155
,
276
Crossover,
186
,
187
,
188
Cross-validation,
257
-
258
,
273
,
291
Books24x7.com, Inc. © 1999-2001 –
Feedback