predict20x {s20x}R Documentation

Model Predictions for a Linear Model

Description

Uses the main output and some error messages from R function "predict" but gives you more output. (Error messages are not reliable when used in Splus.)

Note: The data frame, newdata, must have the same column order and data types (e.g. numeric or factor) as those used in fitting the model.

Usage

predict20x(object, newdata, cilevel = 0.95, digit = 3, print.out = TRUE,
           ...)

Arguments

object an lm object, i.e. the output from "lm()".
newdata prediction data frame.
cilevel confidence level of the interval.
digit decimal numbers after the point.
print.out if T, print out the prediction matrix.
... optional arguments that are passed to the generic "predict"

Value

frame vector or matrix including predicted values, confidence intervals and predicted intervals.
fit prediction values.
se.fit standard error of predictions.
residual.scale residual standard deviations.
df degrees of freedom for residual.
cilevel confidence level of the interval.

See Also

"predict", "predict.lm", "as.data.frame".

Examples

# Zoo data
data(zoo.df)
attach(zoo.df)
day.type<-factor(day.type)
fit1<-lm(log(attendance) ~ time + sun.yesterday + nice.day + day.type + tv.ads)
pred.zoo<-data.frame(8,10.8,0,factor(3),1.181)
predict20x(fit1,pred.zoo)
detach(zoo.df)

# Peruvian Indians data
data(peru.df)
attach(peru.df)
fit2<-lm(BP ~ age + years + I(years^2) + weight + height)
pred.peru<-data.frame(21,2,2^2,71,1629)
predict20x(fit2,pred.peru)
detach(peru.df)

[Package s20x version 3.1-5 Index]