The KendSPear.lng Model


View the model
Download the model

Calculate Kendall tau, Spearman, and Pearson correlation coefficients.
The advantage of Kendall and Spearman is that
they are non-parametric. E.g.,
if you compute the Kendall or Spearman correlations for
a set of uniform random variables, and then
transform these uniforms into Normal random
variables using monotonic increasing transformations,
the Kendall and Spearman correlations remain unchanged.
The values are in the range [-1,+1].
Also, if you are contemplating whether a Normal vs
a LogNormal vs a Poisson distribution best fits the
joint demand distribution for a set of 2 or more products,
the same rank order correlation applies for all three
distribution choices.
For the Normal distribution, the relationship
between the expected standard Pearson linear correlation, rho,
and tau is: rho = sin(tau*pi/2). This relationship
holds more generally for any so-called elliptic distribution;


Correlation Coefficient | Kendall Tau | Non-parametric Statistic | Spearman Rank | Pearson Correlation | Rank Order Statistics |