Kolmogorov–Smirnov test
Prueba de Kolmogórov-Smirnov
- Date
- 1933 · confidence high
- Place
- Moscow, Russia (USSR)
- Credited to
- Andrey Kolmogorov and Nikolai Smirnov
- Field
- Knowledge & symbols
What was discovered
A non-parametric statistical test that compares a sample against a reference distribution, or two samples against each other, by measuring the largest distance between their distribution functions. It is one of the most widely used goodness-of-fit tests.
Kolmogorov formulated the one-sample case in 1933; Smirnov extended the theory to the two-sample case in 1939.
What it led to
- A de facto standard for non-parametric goodness-of-fit testing
- Extensions such as the two-sample Kolmogorov-Smirnov test