compare.estimate.and.truth {dmt}R Documentation

Measuring model accuracy.

Description

Compare estimated parameters and the original parameters used to generate the toydata.

Usage

compare.estimate.and.truth(res, toy)

Arguments

res

DependencyModel object.

toy

Toydata which was used to learn the model.

Details

Given original data (including the original parameters) and model learned based on the data, compare the learned parameters to the original ones.

Value

Vector with following elements

wtw.x, wtw.y

Correlation between the original and estimated values for W*t(W) for X and Y, respectively.

phi.x, phi.y

Correlation between the original and estimated values for phi for X and Y, respectively.

Note

Additional tests added later.

Author(s)

Leo Lahti leo.lahti@iki.fi

References

See citation("dmt") for references.

See Also

generate.toydata

Examples

 
# toy <- generate.toydata(N = N, 
#           zDim = zdim, xDim = xdim, yDim = ydim,
#           marginal.covariances = marginalCovariances,
#           priors = priors)
# res <- fit.dependency.model(toy$X, toy$Y,
#                      zDimension = zdim,
#              marginalCovariances = marginalCovariances,
#		      priors = priors,
#                     matched = FALSE)
# vec <- compare.estimate.and.truth(res, toy)
#

[Package dmt version 0.8.20 Index]