Leon wrote:
> Hi,
> I am using rpart to do leave one out cross validation, but met some
problem,
> Data is a data frame, the first column is the subject id, the second column
is the group id, and the rest columns are numerical variables,
>
>
>>Data[1:5,1:10]
>
> sub.id group.id X3262.345 X3277.402 X3369.036 X3439.895 X3886.935
X3939.054 X3953.777 X3970.352
> 1 32613 HAM_TSP 417.7082 430.4895 619.4776 720.8246 1048.1290
374.4598 770.4653 646.8712
> 2 32616 HAM_TSP 2553.5913 1113.4997 225.5340 274.5644 1105.7908
363.4623 2380.1872 784.9196
> 3 32617 HAM_TSP 291.6596 314.7322 238.3287 315.8305 982.6276
299.5855 1059.1140 540.8592
> 4 32628 HAM_TSP 456.9504 508.2278 552.3632 719.9989 1306.6299
446.6184 1352.9955 867.4219
> 5 32629 HAM_TSP 898.8879 640.2680 342.5477 386.5816 811.6709
518.0244 715.9886 441.1622
>
> Example, I use the first sample as test set, the rest as training set
>
>
>>fit <- rpart(as.factor(Data[-1,2]) ~., Data[-1, -c(1:2 ) ],
minbucket=2 )
>
>
>>predict(fit, Data[1,],type='prob')
>
> Error in predict.rpart(fit, Data[1, ]) : subscript out of bounds
rpart has only seen one class so far, hence the tree consists of the
root only.
Uwe Ligges
> but when I changes the parameter of type into 'class'
> it works well
>
>
>>predict(fit, Data[1,-c(1:2)],type='class')
>
> [1] HTLV_Carrier
> Levels: HAM_TSP HTLV_Carrier Leukemia Normal
>
> Could anyone tell me what is the problem with it?
>
> Thanks very much in advance!
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html