This project aims to conduct A/B testing on the results of an experiment that includes users categorized into Control and Variant groups, along with information on the generated revenue. The dataset was obtained from Kaggle.
The objective is to use appropriate tests for each case, whether it is for mean differences or proportions, even if the data does not strictly meet the assumptions of some tests. This allows us to compare the results and determine when these tests can still be effective despite such conditions.
The tests used in this project were:
- Permutation test
- T-test
- Mann-Whitney U rank test
- Z-test
- Chi-Square test
- Fisher's Exact test