predict.LSBP_VB: Predict method for the LSBP

In blindedmanuscript/LSBP: Dependent Logit Stick-Breaking Process

Description

Predict method for the LSBP estimated using the Variational Bayes algorithm.

Usage

## S3 method for class 'LSBP_VB'
predict(object, type = "mean", R = 5000, newdata = NULL,
  threshold = NULL, ...)

Arguments

object	An object of class LSBP_VB.
type	String indicating the type of prediction: type="mean",type="predictive", type="variance" or type="cdf". See details.
R	An integer indicating the number of replications for the returned sample.
newdata	A new data frame containing the same variables declared in Formula. If missing, the dataset provided for estimation is used.
threshold	Only needed if type="cdf" is selected. See details.
...	Further arguments passed to or from other methods.

Details

The method predict.LSBP_VB produces a sample of predicted values, obtained by evaluating the conditional mean of the LSBP model or the predictive distribution, using the observations contained in the newdata data frame.

If type="mean" a sample from the posterior distribution of the LSBP mean is returned. If type="predictive" is selected, then a sample from the predictive distribution is returned. If type="variance" a sample from the posterior distribution of the LSBP variance is returned. If type="cdf" a sample from the posterior distribution of the LSBP cumulative distribution function is returned, evaluated at threshold.

blindedmanuscript/LSBP documentation built on May 13, 2019, 8:23 a.m.