A better solution provide by Adriano Rivolli:
library(parallel)
models <- mclapply(1:10, function(i) {
model <- RWeka::J48(Species ~., iris[sample(nrow(iris))[1:100],])
rJava::.jcache(model$classifier)
model
});
2013-10-24 10:16 GMT-02:00 Luís Paulo F. Garcia <lpfgar...@gmail.com>:
> Dear, Patrick.
>
> I'm using a workaround to work with RWeka and multicore package. The most
> important point is not loading the RWeka package and use the namespace in a
> encapsulated function. Take a look on this code:
>
>
> # C4.5 classifier. Return the prediction for a test dataset.
> cl.c45 = function(tran, test) {
>
> model = RWeka::J48(Species ~ ., tran);
> pred = predict(model, test[,-ncol(test)], type="class");
> names(pred) = row.names(test);
> return(pred);
> }
>
>
> library(multicore)
> aux = mclapply(1:1000, function(i) {
> cl.c45(iris, iris)
> })
>
> Regards,
> Luis
>
>
>
>
>
>
>
>
> On Tue, Oct 22, 2013 at 7:11 AM, Patrick Connolly <
> p_conno...@slingshot.co.nz> wrote:
>
>> On Thu, 17-Oct-2013 at 02:21PM -0300, Luís Paulo F. Garcia wrote:
>>
>> |> I work very mutch with the packages RWeka and multicore. If you try to
>> run
>> |> J48 or any tree of RWeka with multicore we hava some errors.
>> |>
>> |> Example I:
>> |>
>> |> library(RWeka);
>> |> library(multicore);
>> |>
>> |> mclapply(1:100, function(i) {
>> |> J48(Species ~., iris);
>> |> });
>> |>
>> |>
>> |> Output: "Error in .jcall(o, \"Ljava/lang/Class;\", \"getClass\") : \n
>> |> java.lang.ClassFormatError: Incompatible magic value 1347093252 in
>> class
>> |> file java/lang/ProcessEnvironment$StringEnvironment\n"
>> |>
>> |>
>> |> Example II:
>> |>
>> |> library(multicore);
>> |>
>> |> mclapply(1:100, function(i) {
>> |> RWeka::J48(Species ~., iris);
>> |> });
>> |>
>> |> Output: Erro em .jcall(x$classifier, "S", "toString") :
>> |> RcallMethod: attempt to call a method of a NULL object.
>> |>
>> |>
>> |> Do you know some way to work with parallel processing and RWeka? I
>> tried
>> |> MPI and SNOW without success.
>>
>> Not much help, but I too have not been able to get parallelling RWeka
>> to work. OTOH, what RWeka can do is very fast compared with, say, gbm
>> (which does work well with mclapply).
>>
>> I suspect that it has something to do with how Java is set up, but I
>> know nothing about setting up Java.
>>
>>
>>
>>
>> |>
>> |> R version 3.0.2 (2013-09-25) -- "Frisbee Sailing"
>> |> Ubuntu 12.04 x64
>> |>
>> |>
>> |> --
>> |> Lu?s Paulo Faina Garcia
>> |> Engenheiro de Computa??o - Universidade de S?o Paulo
>> |> S?o Carlos - SP - Brasil
>> |>
>> |> [[alternative HTML version deleted]]
>> |>
>>
>> |> ______________________________________________
>> |> R-help@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.
>>
>>
>> --
>> ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.
>> ___ Patrick Connolly
>> {~._.~} Great minds discuss ideas
>> _( Y )_ Average minds discuss events
>> (:_~*~_:) Small minds discuss people
>> (_)-(_) ..... Eleanor Roosevelt
>>
>> ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.
>>
>
>
>
> --
> Luís Paulo Faina Garcia
> Engenheiro de Computação - Universidade de São Paulo
> São Carlos - SP - Brasil
>
--
Luís Paulo Faina Garcia
Engenheiro de Computação - Universidade de São Paulo
São Carlos - SP - Brasil
[[alternative HTML version deleted]]
______________________________________________
R-help@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.
