Which calculation method is used for Johnson Fit Curves in TIBCO Statistica?

Which calculation method is used for Johnson Fit Curves in TIBCO Statistica?

book

Article ID: KB0075272

calendar_today

Updated On:

Products Versions
Spotfire Statistica All versions

Description

Statistica use two methods for calculating Johnson Fit Curves - using the method of moments (default option), and using the percentile method. The 'JohnsonFromMoments' provides the Probability Density Function (PDF) value for a given x, that is, it computes f(x) where f is the JohnsonDistribution specified by the given moments. The vJohnson is the inverse cumulative Johnson or quantile function for the Johnson distribution specified by the given moments (it provides the x for a given F(x) where F is the Cumulative Distribution Function (CDF) of the specified Johnson distribution).

Issue/Introduction

This article details how Statistica calculates Johnson Fit Curves

Resolution

  1. By default, Johnson Curves in Statistica are fit via the method of moments (which is described in further detail here: https://rss.onlinelibrary.wiley.com/doi/10.2307/2346692 Hill, I. D., Hill, R., & Holder, R. L. (1976) - Fitting Johnson curves by moments. Applied Statistics. 25, 190-192). 
  2. There is a second method that can be used, which is the percentile method ( described in further detail here: Transforming non-normal data to normality in statistical process control. Journal of Quality Technology, 30(2), 133-141 Reference: https://www.tandfonline.com/doi/abs/10.1080/00224065.1998.11979832)   

Also, using Statistica Visual Basic (SVB), you can fit via moments by using 'FitJohnsonFromMoments'