PSGD SVM
mvn clean install
Make sure resource/datasouce.yaml has the base path for all your datasets Each dataset will have it's own folder.
For example epsilon
And within each folder there will be training.csv file which is a dense matrix The format of the file must be as follows.
label, feature_1, feature_2, ...., feature_d
Currently we only support Single file based data splitting and run.
Dynamic run modes will be available in the first release.
Refer to the bin folder scripts.