-
Notifications
You must be signed in to change notification settings - Fork 8.3k
Add statistical aggregate function kolmogorovSmirnovTest #48228
Description
Use case
This type of statistical test is useful to check the equality of two distributions. In addition it can be easily modified to check whether a sample comes from normal distribution by performing the test against a normally distributed sample. (Normal distribution is just an example, you can perform a test against sample from any distribution).
Previously there was an attempt #37873 to add a Shapiro-Wilk test for normality, but it has one very big disadvantage - it works only for relatively small samples (< 2500), which it not applicable for ClickHouse use-case.
Describe the solution you'd like
Take a look how welchTTest or studentTTest are implemented.
Describe alternatives you've considered
As an alternative andersonDarling test could be also implemented.