Combining Pattern Classifiers: Methods and Algorithms, 2nd Edition

INDEX

 

• AdaBoost
  • reweighting and resampling
  • AdaBoost.M1
  • AdaBoost.M2
  • face detection
  • training error bound
  • variants of
• Arcing
• Attribute

 

• Bagging
  • nice
  • out-of-bag data
  • pasting small votes
• Bayes decision theory
• Bayesian learning
• Bias and variance
• Bonferroni-Dunn correction
• Bootstrap sampling

 

• Class
  • labels
    • soft
    • uncertain
  • prevalence
  • separability
• Classes
  • equiprobable
  • linearly separable
  • overlapping
  • unbalanced
• Classification
  • boundaries
  • accuracy
  • margin
  • regions
• Classifier
  • =inducer
  • =learner
  • base
  • canonical model
  • comparison
  • complexity
  • decision tree
  • k-nearest neighbor (k-nn)
    • prototype
    • reference set
  • largest prior
  • linear discriminant (LDC)
    • regularization of
  • Naïve Bayes
  • nearest mean (NMC)
  • neural networks
  • non-metric
  • output
    • abstract level
    • correlation
    • independent
    • measurement level
    • oracle
  • output calibration
  • performance of
  • quadratic discriminant (QDC)
  • selection
  • support vector machine (SVM)
  • unstable
• Classifier competence
  • direct k-nn estimate
  • distance-based k-nn estimate
  • map
  • potential functions
  • pre-estimated regions
• Classifier selection
  • cascade
  • clustering and selection
  • dynamic
  • local class accuracy
  • regions of competence
• Clustering
  • hierarchical
  • k-means
  • non-hierarchical
  • single linkage
    • chain effect
• Combiner
  • average
  • Behavior Knowledge Space (BKS)
  • competition jury
  • decision templates
  • equivalence of
  • generalized mean
    • level of optimism
  • linear regression
  • majority vote
  • median
  • minimum/maximum
  • multinomial
  • Naïve Bayes
  • non-trainable
  • optimality
  • oracle
  • plurality ...