Kolmogorov-Smirnov (K-S) statistic

In my previous blog we read about “Statistics importance in Regression Modeling”. In this blog we will understand one of the concept which is call KS Statistics or K-S statistics.

It is named after Andrey Kolmogorov and Nikolai Smirnov

• The Kolmogorov-Smirnov (K-S) statistic is a non-parametric test used to compare a sample with a reference probability distribution or to compare two samples. It can be used to check the goodness of fit of a distribution to data.
• In this context, we can use the K-S test to check whether the residuals (difference between actual and predicted values) of our regression model follow a normal distribution

Did not understand it yet?

Lets understand through a story

Imagine you have a big jar of jellybeans. You think all the jellybeans in this jar are blue. To check, you close your eyes and take a small handful of jellybeans. Now, you want to compare the colors in your hand to what you expect (all blue).

The Kolmogorov-Smirnov (let’s call it K-S) test is like a tool that helps you decide if the handful of jellybeans you picked is a good representation of the entire jar.

Using the K-S test, you compare your handful to the expected all-blue jellybeans. If the test says, “Hey, this…