Learn R Programming
RSSL (version 0.9.3)

ICLinearDiscriminantClassifier: Implicitly Constrained Semi-supervised Linear Discriminant Classifier

Description

Semi-supervised version of Linear Discriminant Analysis using implicit constraints as described in (Krijthe & Loog 2014). This method finds the soft labeling of the unlabeled objects, whose resulting LDA solution gives the highest log-likelihood when evaluated on the labeled objects only. See also ICLeastSquaresClassifier.

Usage

ICLinearDiscriminantClassifier(X, y, X_u, prior = NULL, scale = FALSE,
  init = NULL, sup_prior = FALSE, x_center = FALSE, ...)

Arguments

X

design matrix of the labeled objects

y

vector with labels

X_u

design matrix of the labeled objects

prior

set a fixed class prior

scale

logical; Should the features be normalized? (default: FALSE)

init

not currently used

sup_prior

logical; use the prior estimates based only on the labeled data, not the imputed labels (default: FALSE)

x_center

logical; Whether the data should be centered

...

Additional Parameters, Not used

References

Krijthe, J.H. & Loog, M., 2014. Implicitly Constrained Semi-Supervised Linear Discriminant Analysis. In International Conference on Pattern Recognition. Stockholm, pp. 3762-3767.

See Also

Other RSSL classifiers: EMLeastSquaresClassifier, EMLinearDiscriminantClassifier, GRFClassifier, ICLeastSquaresClassifier, KernelLeastSquaresClassifier, LaplacianKernelLeastSquaresClassifier(), LaplacianSVM, LeastSquaresClassifier, LinearDiscriminantClassifier, LinearSVM, LinearTSVM(), LogisticLossClassifier, LogisticRegression, MCLinearDiscriminantClassifier, MCNearestMeanClassifier, MCPLDA, MajorityClassClassifier, NearestMeanClassifier, QuadraticDiscriminantClassifier, S4VM, SVM, SelfLearning, TSVM, USMLeastSquaresClassifier, WellSVM, svmlin()