## Hypotheses on true sample distribution

In order to forecast the behavior of the data sample, you need to know not only its parameters, such as mean and variance, but also the distribution law, which controls the data. There are many distribution laws, and to suggest a hypothesis on similarity, you need to know the unique characteristics of each distribution. It is often sufficient to study a sample histogram to make a choice.

Using the `DistributionFitTest`

function, you can test the hypothesis that the dataset was drawn from a population with a distribution, and the alternative hypothesis H_{A} that it was not.

In order to check the main hypothesis, the data sample is tested. This tests the mean assessment of the difference *d(x)* of the empirical value of the distribution function and its predicted value, *F(x)*. The following tests are conducted for univariate or multivariate distributions:

Test's name |
Type of test |
Description |
---|---|---|

"AndersonDarling" |
Distribution, data |
This is based on Expectation... |