quick start

Use my packet byas_ab.py like in quick_start notebooks

bayesian-a-b-testing; Simple pipline: notebooks

p.s this is a project for learning. Supplemented

---

Prior distribution - (Befor experiments) the distribution is taken as an a priori distribution is - Beta(1,1);

Posterior distribution - need to calculate (step 1.)

1. Generate test data;

Control group - Data about the system on which nothing has changed (the website page with the 'buy' button). number of observations, number of target events;
Test group - Data about the same system on which we are testing new hypotheses (changed the color of the "buy" button). number of observations, number of target events;

2. PDF's for test and control group (Calculate posterior probability)

λ - conversion, n - number of observations, c - number of target events, (a,b) - parameters of a priori probability.

3. Calculating the lift

4. Calculating probabiliti - code

References

• Bayesian A/B Testing with Python: the easy guide
• Formulas for Bayesian A/B Testing
• Bayesian A/B Testing: a step-by-step guide