machine-learninglibsvmshogun

Libsvm file format for Shogun

I'm new to shogun and I've been told that it's efficient with large datasets. I keep reading that Shogun supports LibSVM data format so I thought it'd be easier to switch.

I noticed that shogun needs training data and labels set separately. In LibSVM's file format they are both contained in one data file. How can I load the exact same data file that I created for LibSVM in Shogun (i.e. without separating data and labels)?

Solution

  • checkout the latest develop branch of shogun toolbox from github. it has now native support for reading libsvm file format.

    for more details check the examples