[R] residuals

[Shawn Way]

I was going through all kinds of torture to figure this out.  An elegant
solution, indeed.

Thank you very much.

Shawn Way, PE
Engineering Manager
sway at tanox.com


-----Original Message-----
From: Liaw, Andy [mailto:andy_liaw at merck.com] 
Sent: Friday, April 16, 2004 9:15 AM
To: 'Shawn Way'
Subject: RE: [R] residuals


Shawn,

This is what I'd do:

lof <- function(x, y) {
    fit1 <- anova(lm(y ~ x))
    fit2 <- summary(aov(y ~ factor(x)))[[1]]
    SSpe <- fit2[["Sum Sq"]][2]
    DFpe <- fit2[["Df"]][2]
    SSresid <- fit1[["Sum Sq"]][2]
    DFresid <- fit1[["Df"]][2]
    SSlof <- SSresid - SSpe
    DFlof <- DFresid - DFpe
    Fstat <- (SSlof / DFlof) / (SSpe / DFpe)
    pval <- pf(Fstat, DFlof, DFpe, lower.tail=FALSE)
    out <- matrix(c(SSlof, SSpe, DFlof, DFpe, Fstat, NA, pval, NA), nrow=2,
                  dimnames=list(c("Lack-of-fit", "Pure Error"),
                                c("Sum Sq", "Df", "F value", "Pr(>F)")))
    class(out) <- c("anova", class(out))
    out
}

On your data, I get:

> lof(data$ref, data$actual)
             Sum Sq Df F value Pr(>F)
Lack-of-fit 0.13441  1  0.7984 0.4374
Pure Error  0.50505  3               

Best,
Andy

> From: Shawn Way
> 
> 
> I'm trying to determine the lack of fit for regression on the
> following:
> 
> data <- data.frame(ref=c(0,50,100,0,50,100),
>                              actual=c(.01,50.9,100.2,.02,49.9,100.1),
>                              level=gl(3,1))
> fit <- lm(actual~ref,data)
> fit.aov <- aov(actual~ref+Error(level),data)
> 
> According to the information I have, the lack of fit for this
> regression is
> the f-ratio between the residuals of the level contribution and the
> residuals within.  I'm trying to get the information from the 
> fit to make
> this ratio of the residual mean squares.
> 
> Any thoughts?
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list 
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
> http://www.R-project.org/posting-guide.html
> 
> 


----------------------------------------------------------------------------
--
Notice:  This e-mail message, together with any attachments,...{{dropped}}