Loading

GRUBBSTEST

gts outlier

The GRUBBSTEST function detects an outlier in a GTS (or a LIST of GTS), by applying a Grubbs’ test.

This test is done under the assumption that the GTS follows an approximately normal distribution.

It tests whether there is exactly a single outlier in a GTS or not. For an iterative version which can detect multiple outliers, use instead ESDTEST.

This function only applies to GTS of type DOUBLE.

Reference

Grubbs, Frank (February 1969). “Procedures for Detecting Outlying Observations in Samples”. Technometrics (Technometrics, Vol. 11, No. 1).

GRUBBSTEST is available since version 1.0.0.

See also

Signatures

Examples

// Macro used to generate an approximately normal distribution using central limit theorem <% RAND RAND RAND RAND RAND RAND + + + + + 3.0 - %> 'normal' STORE // we create a GTS with an approximately normal distribution NEWGTS 1 100 <% NaN NaN NaN @normal ADDVALUE %> FOR // we add an outlier (> 3.0) 12 NaN NaN NaN 3.001 ADDVALUE DEDUP // we call GRUBBSTEST F GRUBBSTEST