[R] mlbench dataset question

David Winsemius dwinsemius at comcast.net
Mon Jul 6 22:42:41 CEST 2009


On Jul 6, 2009, at 3:16 PM, spime wrote:

>
> Dear R-users,
>
> Recently, I am facing some problems when converting mlbench data  
> into matrix
> format.
>
> library(mlbench)
> data(BostonHousing)
> X<- BostonHousing[,1:13]

If you look at X with str(), you will see it has a factor variable.

> y<-BostonHousing[,14]
>
> I want to convert X and y into matrix form. I am getting these obvious
> errors...
>
>> t(X)%*%y
> Error in t(X) %*% y : requires numeric/complex matrix/vector arguments

And since one of the variables was a factor, the result of t(X) (and   
of as.matrix(X) for that matter ) becomes a character matrix. Perhaps  
data.matrix(X) will serve better?

t(data.matrix(X))%*%y
                [,1]
crim      25687.104
zn       168607.800
indus    111564.080
chas      12397.000
nox        6094.428
rm        73924.078
age      732581.140
dis       45713.874
rad       93445.100
tax     4287797.600
ptratio  205316.730
b       4208049.505
lstat    119799.159

>> t(as.matrix(X))%*%(as.matrix(y))
> Error in t(as.matrix(X)) %*% (as.matrix(y)) :
>  requires numeric/complex matrix/vector arguments
>
> any kind of suggestions will be very helpful.
> Thanks.
> -- 

David Winsemius, MD
Heritage Laboratories
West Hartford, CT




More information about the R-help mailing list