ogc_categoricalResponse: Variational Bayesian inference for outcome-guided clustering...
In acabassi/variational-ogc: Variational Outcome-Guided Clustering

Description Usage Arguments Value References

View source: R/ogc_categorical_response.R

This function allows you do semi-supervised clustering using variational Bayesian inference, when the response variable y is discrete.

1
2
3

ogc_categoricalResponse(X, K, y, prior, R = 0, ms = F, vs = F,
  labels_init, kmeans_init = F, tol = 1e-19, maxiter = 2000,
  verbose = F, indep = T)

`X`	NxD data matrix
`K`	(Maximum) number of clusters
`y`	Vector of response variables of length N.
`prior`	Prior hyperparameters (optional).
`R`	Number of categories in y.
`ms`	Boolean flag that indicates whether model selection is required or not. Default is FALSE.
`vs`	Boolean flag that indicates whether variable selection is required or not. Default is FALSE.
`labels_init`	User-defined initial cluster labels (can be empty).
`kmeans_init`	Boolean flag, which, if TRUE, initializes the cluster labels with the k-means algorithm. Default is FALSE.
`tol`	Tolerance for stopping criterion. Default is 10e-20.
`maxiter`	Maximum number of iterations. Default is 2000.
`verbose`	Boolean flag which, if TRUE, prints the iteration number. Default is FALSE.
`indep`	Boolean flag which determines whether the covariates are considered as independent or not. Default is TRUE.

A list containing L, the lower bound at each step of the algorithm, label, a vector containing the cluster labels, model, a list containing the trained model structure, and a vector called n_ comp which, if model selection is required, contains the number of mixture components at every step of the VB algorithm.

Pattern Recognition and Machine Learning by Christopher M. Bishop

acabassi/variational-ogc documentation built on May 23, 2019, 2:45 p.m.