I would like to train my cluster(clustream in my case) on a specific csv data and then test it on a diffrent csv test data.
here is my code:
dsd <- DSD_ReadCSV("TRAIN.csv", k=NA, take=NULL, class=NULL, loop=FALSE,
sep=",", header=TRUE, skip=0, colClasses = c("NULL",rep(NA,55)))
# pay attention that t1 is a parameter set to 1
t1 <- 1
clustream <- DSC_CluStream(t=t1, m=100, k=23, horizon=500)
update(clustream, dsd, 55)
the update(clustream, dsd, 55) not working well, this is the way to train the cluster on a data? and how could i test it? the meaning of "test it" is to get eventually the cluster id for each record in the test set.
I also came across same problem for different data set, can i know the solution
stream <- DSD_ReadCSV("2c9b0cf998ef4370817f9f41d64fc0a0.csv",header = TRUE,sep=",") clustream <- DSC_CluStream(m=50) update(clustream, stream, 500) Error in .jcall("StreamMOA", "V", "update", object$javaObj,.jarray(as.matrix(d), :
method update with signature (Lmoa/clusterers/AbstractClusterer;[[Ljava.lang.String;)V not found