I used to be in academia, where I used Stan extensively and loved it. I’ve since moved to industry where, for various reasons, I haven’t “needed” Stan. But I miss it! Now, my company is part of a start-up community where we share knowledge, and it occurs to me that while my data crunching doesn’t need/fit-with the rigor of full Bayesian inference, others here might. So I’m planning a talk on “Bayes for Business”, but being new to the non-academic/non-niche-modelling world, I don’t have much personal experience with the kinds a data/inference scenarios that might be encountered by businesses. I’ve sent a query to the local community for examples, but I thought I’d also ask here in case anyone already has experience with benefits earned by moving from more limited tools to full Bayes in a business/industry context. Thanks in advance for your suggestions!
5 Likes
Obviously there are a lot of applications of MCMC in Finance. You’re likely already aware of @Risk and Crystal Ball. Maybe you’ve seen this: https://www.probabilitymanagement.org/. This is also interesting: https://www.getguesstimate.com/.
3 Likes
See Productization of Stan tutorial and panel in StanCon 2019 Helsinki
4 Likes
In pharma we use informative priors derived from literature data to reduce the control group sample size, for example. These approaches are used in early drug development (phase I+II).
I developed the R package RBesT (https://cran.r-project.org/package=RBesT) to do just that…have a read of the “Getting started …” vignette to get an idea.
3 Likes
What is sun?