search

UBC-Stat-ML / bayonet

Probabilistic inference utils
BSD 2-Clause "Simplified" License
4 stars 4 forks source link

Building factors #5

Closed jewellsean closed 10 years ago

jewellsean commented 10 years ago

This question relates to how Blang is intended to be extended to different (non-abstract) factors. In particular, would Blang ship with a variety of ready-to-use factors for, say, trees in phylogenetics or DPMM problems? Or, should each particular application (eg. conifer) have Blang's abstract factors and other interfaces implemented?

alexandrebouchard commented 10 years ago

I was thinking to put in the core project only well known, generic distribution (e.g. univariate ones); and to have one separate project for BNP (namely, Polya), and one for phylo (namely, conifer).