GraphCRF model for a graph without any edges

47 views

Skip to first unread message

amila silva

unread,

Mar 11, 2018, 11:33:04 PM3/11/18

to pystruct

Hi,

Is there anyway to train graphCRF model for a graph without any edges (simply for an isolated set of nodes)?

I just want check the effectiveness of having edges and infer via LBP.

How can I see the values of the edge potentials after the training process? Because, the w attribute of the learners outputs the parameters of the joint features and I'm not clear about how this joint feature vector is created.