1.1 - 8.10 - Distribution Matching (ML Engine) - Teradata Vantage

Teradata Vantage™ - Machine Learning Engine Analytic Function Reference

Teradata Vantage
Release Number
October 2019
Content Type
Programming Reference
Publication ID
English (United States)

Given sample data and reference distributions, the function tests the hypothesis that the sample data comes from the distributions (Hypothesis-Test Mode (ML Engine)). Given the test results, the function finds the distribution that best matches the sample data (Best-Match Mode (ML Engine)).

The Distribution Matching function is composed of the functions DistributionMatchReduce and DistributionMatchMultiInput. DistributionMatchReduce supports these distributions:

  • For continuous variables:
    • Beta
    • Cauchy
    • ChiSq
    • Exponential
    • F
    • Gamma
    • Lognormal
    • Normal
    • T
    • Triangular
    • Uniform
    • Weibull
  • For discrete variables:
    • Binomial
    • Geometric
    • Negative binomial
    • Poisson
    • Uniform

For evaluating the fit of the distribution to the data, the function supports these tests:

  • Anderson-Darling test
  • Kolmogorov-Smirnov test
  • Cramér-von Mises criterion (hypothesis testing only)
  • Pearson’s Chi-squared test