[R] Need help to locate my mistake

Louise Hoffman louise.hoffman at gmail.com
Sun Mar 2 21:18:29 CET 2008


Dear readers

I would like to make General Linear Model (GLM) for the following data set
http://louise.hoffman.googlepages.com/fuel.csv

The code I have written is

fuelData<-read.table('fuel.csv',header=TRUE, sep=',')
n<-dim(fuelData)[1]
xOnes<- matrix(1,nrow=n,ncol=1)
x<-cbind(xOnes,fuelData[,3])
y<-fuelData[,4]
theta<-((t(x)%*%x)^(-1))%*%t(x)%*%y

which gives

> theta
            [,1]
[1,] 215.8374077
[2,]   0.1083491

When I do it in Matlab I get theta to be
79.69
0.18

which is correct. ~79 is the crossing of the y-axis.

Have I perhaps written theta wrong? The formula for theta is
(alpha,beta)^T = (x^T * x)^(-1) * x^T * Y

where ^T means transposed.

Can someone see where the mistake is?

Hugs,
Louise



More information about the R-help mailing list