How to use this about pyds

[sporkmonger]

Generate a set of masses for each classification source, use combination function as per examples, then do pignistic conversion. Whichever is highest is your final classification.

Generating your masses is potentially the hard part. Unless you've built from ground up for DS, you are more likely to have data in simple Bayesian form (I.e. all singleton masses add up to 1.0).