Addz {NCSampling}

Addz

Description

Add variable/s to the population file which are good predictors of the variables/s of interest

Usage

Addz(popfile, training, yvars, xvars, pool)

Arguments

popfile

dataframe containing population data - as a minimum there must be columns named 'PID' (plot identifier), 'Strata' and 'plot_type'.

training

dataframe containing training data. Must contain auxiliary variables and variable/s of interest.

yvars

vector containing the name of each variable of interest (dependent variable).

xvars

vector containing the names of the auxiliary variables.

pool

logical value - should the training data be pooled across strata prior to fitting the regression model?

Details

The predictor variable for the each variable of interest (dependent variable) is obtained by performing random forest regression on the training data using the designated auxiliary variables. The training data can be pooled across strata (pool=T), or fitted separately within each strata (the default). Not normally called directly.

Value

A list with components:-

popfile

population file - data frame, as above, with predictor variable/s added to the file

r.sqared

dataframe containing the R-squared values obtained from the random forest regression/s

Author(s)

G. Melville

References

Random forest regression is performed using the randomForest package.

See Also

DesVar, randomForest.

Examples

## Addz(popfile, training, yvars, xvars)