Skip to contents

Predictions from an ols fit

Source: R/predict.R
predict.ols.Rd

Returns predictions for new data from a fitted ols model, with optional standard errors and exact closed-form confidence or prediction intervals.

Usage

# S3 method for class 'ols'
predict(
  object,
  newdata = NULL,
  type = "response",
  se.fit = FALSE,
  interval = c("none", "confidence", "prediction"),
  level = 0.95,
  robust = c("none", "HC0", "HC1", "HC2", "HC3"),
  ...
)

Arguments

object

An object of class "ols".

newdata

A numeric matrix or data frame of new predictors. NULL (default) returns object$fitted.values.

type

Prediction scale. Only "response" is supported (the linear predictor coincides with the response for a gaussian fit).

se.fit

If TRUE, the returned vector carries an "se.fit" attribute holding the per-row standard error of the fitted mean. For bagged fits the standard error is the per-row bag standard deviation. Default FALSE.

interval

Type of interval to return.

  • "none" (default): point predictions only.

  • "confidence": a matrix with columns c("fit", "lwr", "upr") for the mean response.

  • "prediction": a matrix with columns c("fit", "lwr", "upr") for a new observation (adds the residual variance, or the variance-model variance when the fit was built with varmod = "const" or "lm").

level

Confidence/prediction level. Default 0.95.

robust

Heteroscedasticity-consistent covariance used for the standard errors / intervals.

  • "none" (default): the classical sigma2 (X'WX)^-1.

  • "HC0", "HC1", "HC2", "HC3": a sandwich estimator.

...

Currently unused.

Value

A numeric vector of predictions, or a matrix with columns c("fit", "lwr", "upr") when interval != "none". With se.fit = TRUE the result carries an "se.fit" attribute.

Examples

fit <- ols(mpg ~ wt + hp, data = mtcars)
predict(fit, mtcars[1:5, ], interval = "confidence")
#>                        fit      lwr      upr
#> Mazda RX4         23.57233 22.45623 24.68843
#> Mazda RX4 Wag     22.58348 21.51622 23.65074
#> Datsun 710        25.27582 23.97441 26.57723
#> Hornet 4 Drive    21.26502 20.10932 22.42072
#> Hornet Sportabout 18.32727 17.30889 19.34564