Dear R-Users,
i want to use the function svm of the e1071 package to predict missing data
############################################################
data(iris)
## create missing completely at random data
for (i in 1:5)
{
mcar <- rbinom(dim(iris)[1], size=1, prob=0.1)
iris[mcar == 1, i] <- NA
}
ok <- complete.cases(iris)
model <- svm(Species ~ ., data=iris[ok,])
## try to predict the missing values for Species
## neither
pred <- predict(model, iris[5])
## nor
pred <- predict(model, iris[!ok, -5])
## seems to work....
Many thanks if anyone could tell me what i do wrong and what is the
problem here.
best regards
Andreas
>>>>> On Mon, 28 Sep 2009 16:12:11 +0200, >>>>> Andreas Wittmann (AW) wrote:That is a bug in predict.svm, I will inform David Meyer, the author of the function. Best, Fritz > Dear R-Users, > i want to use the function svm of the e1071 package to predict missing data > ############################################################ > data(iris) > ## create missing completely at random data > for (i in 1:5) > { > mcar <- rbinom(dim(iris)[1], size=1, prob=0.1) > iris[mcar == 1, i] <- NA > } > ok <- complete.cases(iris) > model <- svm(Species ~ ., data=iris[ok,]) > ## try to predict the missing values for Species > ## neither > pred <- predict(model, iris[5]) > ## nor > pred <- predict(model, iris[!ok, -5]) > ## seems to work.... > Many thanks if anyone could tell me what i do wrong and what is the > problem here. > best regards > Andreas
Hi Andreas, Andreas Wittmann wrote:> Dear R-Users, > > i want to use the function svm of the e1071 package to predict missing data > > ############################################################ > > data(iris) > > ## create missing completely at random data > for (i in 1:5) > { > mcar <- rbinom(dim(iris)[1], size=1, prob=0.1) > iris[mcar == 1, i] <- NA > } > > ok <- complete.cases(iris) > > model <- svm(Species ~ ., data=iris[ok,]) > > ## try to predict the missing values for Species > ## neither > pred <- predict(model, iris[5]) > ## nor > pred <- predict(model, iris[!ok, -5]) > ## seems to work....ind <- is.na(iris[,5]) & !apply(iris[,-5], 1, function(x) any(is.na(x)) predict(model, iris[ind,-5]) Best, Jim> > Many thanks if anyone could tell me what i do wrong and what is the > problem here. > > best regards > > Andreas > > ______________________________________________ > 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.-- James W. MacDonald, M.S. Biostatistician Douglas Lab University of Michigan Department of Human Genetics 5912 Buhl 1241 E. Catherine St. Ann Arbor MI 48109-5618 734-615-7826