Page 413 - Understanding Machine Learning
P. 413
Index
3-term DNF, 79 confidence, 18, 22
F 1 -score, 207 consistency, 66
1 norm, 149, 286, 315, 335 Consistent, 247
contraction lemma, 331
accuracy, 18, 22 convex, 124
activation function, 229 function, 125
AdaBoost, 101, 105, 314 set, 124
all-pairs, 191, 353 strongly convex, 140, 160
approximation error, 37, 40 convex-Lipschitz-bounded learning, 133
auto-encoders, 319 convex-smooth-bounded learning, 133
covering numbers, 337
backpropagation, 237 curse of dimensionality, 224
backward elimination, 314
bag-of-words, 173 decision stumps, 103, 104
base hypothesis, 108 decision trees, 212
Bayes optimal, 24, 30, 221 dendrogram, 266, 267
Bayes rule, 306 dictionary learning, 319
Bayesian reasoning, 305 differential set, 154
Bennet’s inequality, 376 dimensionality reduction, 278
Bernstein’s inequality, 376 discretization trick, 34
bias, 16, 37, 40 discriminative, 295
bias-complexity tradeoff, 41 distribution free, 295
Boolean conjunctions, 29, 54, 78 domain, 13
boosting, 101 domain of examples, 26
boosting the confidence, 112 doubly stochastic matrix, 205
boundedness, 133 duality, 176
strong duality, 176
C4.5, 215 weak duality, 176
CART, 216 Dudley classes, 56
chaining, 338
Chebyshev’s inequality, 373 efficient computable, 73
Chernoff bounds, 373 EM, 301
class-sensitive feature mapping, 193 Empirical Risk Minimization, see ERM
classifier, 14 empirical error, 15
clustering, 264 empirical risk, 15, 27
spectral, 271 entropy, 298
compressed sensing, 285 relative entropy, 298
compression bounds, 359 epigraph, 125
compression scheme, 360 ERM, 15
computational complexity, 73 error decomposition, 40, 135
395
