Hi, I'm running into the same issue. I run is.factor() on training data and validation data. All returned false. -- View this message in context: http://r.789695.n4.nabble.com/Where-command-in-ctree-party-tp3264187p4397640.html Sent from the R help mailing list archive at Nabble.com.
"Where" command in ctree (party)
5 messages · Achim Zeileis, Joseph Wang
i run class(). Both have the same class. -- View this message in context: http://r.789695.n4.nabble.com/Where-command-in-ctree-party-tp3264187p4397693.html Sent from the R help mailing list archive at Nabble.com.
On Fri, 17 Feb 2012, josephw wrote:
i run class(). Both have the same class.
Without a reproducible example it is hard to say what is going on here. All columns of the learning and the test data need to have (a) the same name, (b) the same class, (c) the same levels in case they are factors.
-- View this message in context: http://r.789695.n4.nabble.com/Where-command-in-ctree-party-tp3264187p4397693.html Sent from the R help mailing list archive at Nabble.com.
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
I was able to narrow down to a column that has different types using
featurefields<-c(7, 17, 19, 20, 22, 33, 35, 36, 38, 44:132)
flag <- rep(0, ncol(data))
for (i in featurefields) {
flag[i] <- class(data[,i]) != class(validdata[,i])
}
It is an integer class in the training data but is numeric in the validation data even though
unique(valid[,88]) shows 0 and 1.
-----Original Message-----
From: Achim Zeileis [mailto:Achim.Zeileis at uibk.ac.at]
Sent: Friday, February 17, 2012 10:51 AM
To: Joseph Wang
Cc: r-help at r-project.org
Subject: Re: [R] "Where" command in ctree (party)
On Fri, 17 Feb 2012, josephw wrote:
i run class(). Both have the same class.
Without a reproducible example it is hard to say what is going on here. All columns of the learning and the test data need to have (a) the same name, (b) the same class, (c) the same levels in case they are factors.
-- View this message in context: http://r.789695.n4.nabble.com/Where-command-in-ctree-party-tp3264187p4 397693.html Sent from the R help mailing list archive at Nabble.com.
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
On Fri, 17 Feb 2012, Joseph Wang wrote:
I was able to narrow down to a column that has different types using
featurefields<-c(7, 17, 19, 20, 22, 33, 35, 36, 38, 44:132)
flag <- rep(0, ncol(data))
for (i in featurefields) {
flag[i] <- class(data[,i]) != class(validdata[,i])
}
A construct like which(sapply(data, class) != sapply(validdata, class)) might also help.
It is an integer class in the training data but is numeric in the validation data even though unique(valid[,88]) shows 0 and 1.
Hopefully valid[,88] <- as.integer(valid[,88]) should then resolve the problems. Z
-----Original Message----- From: Achim Zeileis [mailto:Achim.Zeileis at uibk.ac.at] Sent: Friday, February 17, 2012 10:51 AM To: Joseph Wang Cc: r-help at r-project.org Subject: Re: [R] "Where" command in ctree (party) On Fri, 17 Feb 2012, josephw wrote:
i run class(). Both have the same class.
Without a reproducible example it is hard to say what is going on here. All columns of the learning and the test data need to have (a) the same name, (b) the same class, (c) the same levels in case they are factors.
-- View this message in context: http://r.789695.n4.nabble.com/Where-command-in-ctree-party-tp3264187p4 397693.html Sent from the R help mailing list archive at Nabble.com.
______________________________________________ R-help at r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.