I’m looking to infer a single parameter, m, for several sets of observations, whereby the inferred parameter will be different for each set. If I have 5 sets of 50 observations each, I am able to run my experiment nicely by setting shape=5 in my prior for m and then making sure my observations and …

Just reshape everything to be 1d arrays and use indexing to identify which “observation set” the observation applies to. However I am completely stumped for how to do this if I have say, 44 input/observations pairs in one set, 35 in another, 51 in another and so on Create an 1D integer array tha…

Passing observations of varying size to a likelihood function

Questions

cwilliam May 22, 2019, 10:45pm 3

Thanks very much - exactly what I’m after.

Topic		Replies	Views
How to deal with a variable size of the 'observed' variable Questions	15	4003	October 16, 2017
Problem model definition PyMC3 Questions	12	4749	September 21, 2017
Making predictions with a different number of data points than used for estimating the parameters Questions	3	753	September 29, 2018
Latent GP prior on X=theano.shared node with changing shape Questions theano	2	636	November 17, 2017
Combining parameters into single tensor Questions	19	3538	October 30, 2017

Passing observations of varying size to a likelihood function

Related topics