This document summarizes a test case from literature that compares the results of multidimensional scaling using PERMAP software to those quoted in a reference. PERMAP is shown to produce a Stress1 value of 0.01717 for a 4 object dataset using euclidean distances, consistent with the value of 0.01739 reported in the reference, validating that PERMAP implements the method correctly.
This document summarizes a test case from literature that compares the results of multidimensional scaling using PERMAP software to those quoted in a reference. PERMAP is shown to produce a Stress1 value of 0.01717 for a 4 object dataset using euclidean distances, consistent with the value of 0.01739 reported in the reference, validating that PERMAP implements the method correctly.
This document summarizes a test case from literature that compares the results of multidimensional scaling using PERMAP software to those quoted in a reference. PERMAP is shown to produce a Stress1 value of 0.01717 for a 4 object dataset using euclidean distances, consistent with the value of 0.01739 reported in the reference, validating that PERMAP implements the method correctly.
Use Euclidean distances, ratio MDS, two dimensions, and Stress1.
This shows that the same half-matrix definition is being used and that the Stress1 formulas are the same as those shown by Borg and Groenen. The literature is not uniform on either of these points. Borg and Groenen got Stress1 = .01739. Permap gets 0.01717.