Application of Bayesian Inference in NoSQL Database Benchmark

We’re currently doing a study about difference NoSQL Databases, and i’m planning to incorporate a statistical analysis. Basically, we’re comparing 3 Databases and we’re testing the following

  • Number of Rows it can load per second
  • Number of Rows it can read per second

I plan on recording that metric 50 times for each DBs. I plan on following this tutorial on Pymc3 website.

My question is,

  • Is it possible to replicate this with multiple groups?
  • I don’t think student-T distribution is applicable here, what should be a more suitable prior?

Thanks :slight_smile:

normal or lognormal distribution (or other distributions that are defined on Real+ as the observed is positive only)?

1 Like

Thanks :slight_smile: I’ll update this post analysis is done.

1 Like