Testing SVM models & trying to predict with diabetes data taken from kaggle.com. There are three SVM models below using 'kernlab', 'pROC' & 'e1071' package via 'caret' package. All three models use same trainControl but different methods, 'svmRadial', 'svmLinearWeights' & 'svmRadialWeights'.
library(caret)
library(kernlab)
library(pROC)
library(e1071)
This is a demonstration on how to run svm with caret package in R. There are three different svm methods used, svmRadial, svmLinearWeights & svmRadialWeights. The methods use same trainControl parameters and then see which of these three methods performs better on kaggle.com data. All three methods done in here, can be executed without using caret package, but I think caret makes it easier and also provides a lot of different resampling methods and other parameters.