python - scipy: Drawing from sampled probability distribution -
i trying draw random samples probability distribution shape given samples of x,y , z data points.
i've read articles rv_continuous function allows create custom probability distribution. also, i've seen gaussian_kde function allows fit gaussian kernel data points.
however, both ways won't work in case. rv_continuous seems require closed form of function don't have. want draw random samples arbitrary 2 dimensional probability distribution has been sampled. thus, data consists of x,y,z data points (or rather meshgrids).
is there way tell scipy should use these 3 data sets , build multi dimensional pdf it? guess internal representation similar or @ least problem rather frequent one. workaround, imagine using rv_continuous , returning value dataset. however, new python , don't know how pass or access data within class.
thanks help!
Comments
Post a Comment