predict.densityMclustBounded {mclustAddons}

Model-based mixture density estimation for bounded data

Description

Compute density estimation for univariate and multivariate bounded data based on Gaussian finite mixture models estimated by densityMclustBounded.

Usage

## S3 method for class 'densityMclustBounded'
predict(object, newdata, 
        what = c("dens", "cdens", "z"), 
        logarithm = FALSE, ...)

Arguments

object

An object of class "densityMclustBounded" resulting from a call to densityMclustBounded.

newdata

A numeric vector, matrix, or data frame of observations. If missing the density is computed for the input data obtained from the call to densityMclustBounded.

what

A character string specifying what to retrieve: "dens" returns a vector of values for the mixture density; "cdens" returns a matrix of component densities for each mixture component (along the columns); "z" returns a matrix of component posterior probabilities.

logarithm

A logical value indicating whether or not the logarithm of the densities/probabilities should be returned.

...

Further arguments passed to or from other methods.

Value

Returns a vector or a matrix of values evaluated at newdata depending on the argument what (see above).

Author(s)

Luca Scrucca

References

Scrucca L. (2019) A transformation-based approach to Gaussian mixture density estimation for bounded data. Biometrical Journal, 61:4, 873–888. https://doi.org/10.1002/bimj.201800174

See Also

densityMclustBounded, plot.densityMclustBounded.

Examples

y <- sample(0:1, size = 200, replace = TRUE, prob = c(0.6, 0.4))
x <- y*rchisq(200, 3) + (1-y)*rchisq(200, 10)
dens <- densityMclustBounded(x, lbound = 0)
summary(dens)
plot(dens, what = "density", data = x, breaks = 11)

xgrid <- seq(0, max(x), length = 201)
densx <- predict(dens, newdata = xgrid, what = "dens")
cdensx <- predict(dens, newdata = xgrid, what = "cdens")
cdensx <- sweep(cdensx, MARGIN = 2, FUN = "*", dens$parameters$pro)
plot(xgrid, densx, type = "l", lwd = 2)
matplot(xgrid, cdensx, type = "l", col = 3:4, lty = 2:3, lwd = 2, add = TRUE)

z <- predict(dens, newdata = xgrid, what = "z")
matplot(xgrid, z, col = 3:4, lty = 2:3, lwd = 2, ylab = "Posterior probabilities")