classification - How to perform cross-validation for SVM in R?

Question

I am trying to build a SVM classifier for medical data (https://drive.google.com/file/d/1lIehAVBzR5B1NHS6-ozvqpJ9TRNK0VJu/view?usp=sharing) but I can't find the reason why it is always predicting recid as NO. It is happening even when the probability of YES is over 50%.

library(e1071)
library(caTools)

df <- read.table('datos_icb.txt',header=TRUE, stringsAsFactors = TRUE)

sample = sample.split(df$recid, SplitRatio = 0.75)
train = subset(df, sample == TRUE)
test  = subset(df, sample == FALSE)

tune.out=tune(svm,recid~.,data=df, 
              ranges=list(cost=c(0.001, 0.01, 0.1, 1,5,10,100), 
                          gamma = c(0.5, 1, 2, 3,4)))
summary(tune.out)
bestmod=tune.out$best.model
summary(bestmod)

ypred <- predict(bestmod, test)
table(predict=ypred, truth=test$recid)

Could it be a data type problem?

Thanks in advance.

question from:https://stackoverflow.com/questions/66061470/how-to-perform-cross-validation-for-svm-in-r

Answer 1

Waitting for answers