predict.wa: Predict from a weighted average model

In analogue: Analogue and Weighted Averaging Methods for Palaeoecology

Predict from a weighted average model

Description

Model predictions and cross-validation predictions for weighted averaging transfer function models.

Usage

## S3 method for class 'wa'
predict(object, newdata,
        CV = c("none", "LOO", "bootstrap", "nfold"),
        verbose = FALSE, n.boot = 100, nfold = 5, ...)

Arguments

object	an object of class "wa", usually the result of a call to wa
newdata	An optional data frame in which to look for variables with which to predict. If omitted, the fitted values are used.
CV	Should cross-validation be performed? Leave-one-out ("LOO"), bootstrap ("bootstrap") and k-fold ("nfold") CV are currently available.
verbose	Should CV progress be printed to the console?
n.boot	The number of bootstrap samples or k-fold steps.
nfold	Number of subsets in k-fold CV.
...	further arguments passed to or from other methods.

Details

Not all CV methods produce the same output. CV = "bootstrap" and CV = "nfold" produce sample specific errors.

Value

An object of class "predict.wa", a list with the following components:

pred	A list with components pred and rmsep containing the predicted values and the sample specific errors if available.
performance	A list with model performance statistics.
model.pred	A list with components pred and rmsep containing the predicted values for the training set samples and the sample specific errors if available.
call	the matched function call.
CV.method	The CV method used.

Author(s)

Gavin L. Simpson and Jari Oksanen (k-fold CV)

References

Birks, H.J.B., Line, J.M., Juggins, S., Stevenson, A.C. and ter Braak, C.J.F. (1990). Diatoms and pH reconstruction. Philosophical Transactions of the Royal Society of London; Series B, 327; 263–278.

See Also

wa, predict.mat, performance, reconPlot.

Examples

## Imbrie and Kipp
data(ImbrieKipp)
ImbrieKipp <- ImbrieKipp / 100
data(SumSST)
ik.wa <- wa(SumSST ~ ., data = ImbrieKipp, tol.dw = TRUE,
            min.tol = 2, small.tol = "min")
ik.wa

## load V12.122 core data
data(V12.122)
V12.122 <- V12.122 / 100

## predict summer sea-surface temperature for V12.122 core
set.seed(2)
v12.pred <- predict(ik.wa, V12.122, CV = "bootstrap", n.boot = 100)

## draw the fitted reconstruction
reconPlot(v12.pred, use.labels = TRUE, display = "bars")

## extract the model performance stats
performance(v12.pred)

analogue documentation built on Sept. 30, 2024, 9:41 a.m.