[R] Fwd: missing in neural network

Charles Determan Jr deter088 at umn.edu
Wed Mar 25 13:22:45 CET 2015


Soheila,

Set the name of the last column and remove the comma indexing the column
names.  It is a vector and therefore doesn't need a comma for indexing.
Also, after loading your dataset I realized you also have invalid column
names.  The mixes between hyphens and underscores makes the as.formula call
unhappy.  A simple way to fix is to just change all hyphens to underscores
with gsub.

# name 'resp'
colnames(mydata)[ncol(mydata)] <- "resp"

# change hyphens to underscores
colnames(mydata) <- gsub("-","_",colnames(mydata))

# create formula (without comma index on column names)
fm <- as.formula(paste("resp ~", paste(colnames(mydata)[1:3110],
collapse="+")))

# call neuralnet
out <- neuralnet(fm,data=mydata, hidden = 4, lifesign = "minimal",
linear.output = FALSE, threshold = 0.1)

Best regards,
Charles


On Wed, Mar 25, 2015 at 3:29 AM, Soheila Khodakarim <lkhodakarim at gmail.com>
wrote:

> Dear Charles,
>
> I rewrote code :
> library("neuralnet")
> resp<-c(1,1,1,0,1,0,1,0,1,0,1,0,1,0,1,1,0,1,0,1))
> mydata <- cbind(data24_2, resp)
> dim(mydata)
>  >  20 3111
> fm <- as.formula(paste("resp ~ ", paste(colnames(mydata)[,1:3110],
> collapse="+")))
> > Error in colnames(mydata)[, 1:3110] : incorrect number of dimensions
> :(((
>
> AND
>
> fm <- as.formula(paste(colnames(mydata)[,3111],
> paste(colnames(mydata)[,1:3110], collapse="+")))
> > Error in colnames(mydata)[, 3111] : incorrect number of dimensions
>
> Best,
> Soheila
>
> On Wed, Mar 25, 2015 at 11:12 AM, Soheila Khodakarim <
> lkhodakarim at gmail.com> wrote:
>
>> Hi Charles,
>> Many thanks for your help. I will check and let you know.
>>
>> Best Wishes,
>> Soheila
>> On Mar 25, 2015 12:17 AM, "Charles Determan Jr" <deter088 at umn.edu> wrote:
>>
>>> Soheila,
>>>
>>> Did my second response help you?  It is polite to close say if so, that
>>> way others who come across the problem no that it was solved.  If not, feel
>>> free to update your question.
>>>
>>> Regards,
>>> Charles
>>>
>>> On Tue, Mar 24, 2015 at 9:58 AM, Soheila Khodakarim <
>>> lkhodakarim at gmail.com> wrote:
>>>
>>>> Dear Charles,
>>>>
>>>> Thanks for your guide.
>>>> I run this code:
>>>>
>>>> library("neuralnet")
>>>> resp<-c(1,1,1,0,1,0,1,0,1,0,1,0,1,0,1,1,0,1,0,1)
>>>> mydata <- cbind(data, resp)
>>>> out1 <- neuralnet(resp~mydata[,1:3110],data=mydata, hidden = 4,
>>>> lifesign =
>>>> "minimal", linear.output = FALSE, threshold = 0.1)
>>>>
>>>> I saw this error
>>>>
>>>> Error in neurons[[i]] %*% weights[[i]] : non-conformable arguments
>>>>
>>>> :(:(:(
>>>>
>>>> What should I do now??
>>>>
>>>> Regards,
>>>> Soheila
>>>>
>>>>
>>>> On Tue, Mar 24, 2015 at 3:48 PM, Charles Determan Jr <deter088 at umn.edu>
>>>> wrote:
>>>>
>>>> > Hi Soheila,
>>>> >
>>>> > You are using the formula argument incorrectly.  The neuralnet
>>>> function
>>>> > has a separate argument for data aptly names 'data'.  You can review
>>>> the
>>>> > arguments by looking at the documentation  with ?neuralnet.
>>>> >
>>>> > As I cannot reproduce your data the following is not tested but I
>>>> think
>>>> > should work for you.
>>>> >
>>>> > # Join your response variable to your data set.
>>>> > mydata <- cbind(data, resp)
>>>> >
>>>> > # Run neuralnet
>>>> > out <- neuralnet(resp ~ ., data=mydata, hidden = 4, lifesign =
>>>> "minimal",
>>>> >                        linear.output = FALSE, threshold = 0.1,na.rm =
>>>> > TRUE)
>>>> >
>>>> >
>>>> > Best,
>>>> > Charles
>>>> >
>>>> > On Tue, Mar 24, 2015 at 4:47 AM, Soheila Khodakarim <
>>>> lkhodakarim at gmail.com
>>>> > > wrote:
>>>> >
>>>> >> Dear All,
>>>> >>
>>>> >> I want to run "neural network" on my dataset.
>>>> >> ##########################################################
>>>> >> resp<-c(1,1,1,0,1,0,1,0,1,0,1,0,1,0,1,1,0,1,0,1)
>>>> >> dim(data)
>>>> >> #20*3110
>>>> >>
>>>> >> out <- neuralnet(y ~ data, hidden = 4, lifesign = "minimal",
>>>> linear.output
>>>> >> = FALSE, threshold = 0.1,na.rm = TRUE)
>>>> >> ################################################################
>>>> >> but I see this Error
>>>> >> Error in varify.variables(data, formula, startweights,
>>>> learningrate.limit,
>>>> >>  :
>>>> >>   argument "data" is missing, with no default
>>>> >>
>>>> >> What should I do now??
>>>> >>
>>>> >> Best Regards,
>>>> >> Soheila
>>>> >>
>>>> >>         [[alternative HTML version deleted]]
>>>> >>
>>>> >> ______________________________________________
>>>> >> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>>>> >> 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.
>>>> >>
>>>> >
>>>> >
>>>> >
>>>> >
>>>> >
>>>>
>>>>         [[alternative HTML version deleted]]
>>>>
>>>> ______________________________________________
>>>> R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>>>> 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.
>>>>
>>>
>>>
>>>
>

	[[alternative HTML version deleted]]



More information about the R-help mailing list