Hi guys, I’m new to pymc3 world but In my opinion this tool is fantastic! I’m working on a simple crowdsensing application where a group of users send me estimates of a target (continuous measurements), for example gps location. In my model I want to find the mean value of these measures and a relia…

I see what you are doing now - to me it makes sense. As for the Bad initial energy problem, I see two potential problem: it is likely related to the jitter initialization that sometimes make some values go outside of the support. Try trace = pm.sample(1000, init='adapt_diag') in your data simula…

Data aggregation in crowdsensing application

Questions

junpenglao April 11, 2018, 8:53am 12

Glad to help but it might be easier if you have some code and data set up already. The recommended workflow is to generated some simulation data and graduatelly build more complex model.

Topic		Replies	Views
Marginal likelihood implementation and measurement uncertainty Questions	12	910	November 2, 2019
Problems with Hierachical Bayesian Model - Advertising Questions	16	1081	August 10, 2018
Sampler occasionally gets static in pymc3.1, but not in pymc3.0 Questions development	19	2271	July 5, 2018
Let's introduce ourselves!	35	5234	February 19, 2019
Aggregation of trace in models? - Probabilities for max p of binomial dists Questions	7	1020	October 31, 2017

Data aggregation in crowdsensing application

Related topics