illustrating Legendre's original geodesic calculation
Quote from colinr on September 29, 2024, 1:52 amI am writing an article on explainable AI and interpretable ML, and want to start with Legendre's geodesic least squares calculation - when 'accuracy' and 'interpretability' were aligned.
The formulae are discussed in chapter one of Stigler's "The history of statistics" (q.v. p.57 of http://www.med.mcgill.ca/epidemiology/Hanley/bios601/GaussianModel/CombiningObservationsStiglerCh1.pdf).
However, I really want to see a diagram 'interpreting' them - an ellipse with the coefficients estimated by Legendre marked on them. (It would be particularly nice to see his estimated formulae constructed on the diagram.)
If anyone had seen a diagram like this, I would be grateful to hear from you.
Best,
Colin
I am writing an article on explainable AI and interpretable ML, and want to start with Legendre's geodesic least squares calculation - when 'accuracy' and 'interpretability' were aligned.
The formulae are discussed in chapter one of Stigler's "The history of statistics" (q.v. p.57 of http://www.med.mcgill.ca/epidemiology/Hanley/bios601/GaussianModel/CombiningObservationsStiglerCh1.pdf).
However, I really want to see a diagram 'interpreting' them - an ellipse with the coefficients estimated by Legendre marked on them. (It would be particularly nice to see his estimated formulae constructed on the diagram.)
If anyone had seen a diagram like this, I would be grateful to hear from you.
Best,
Colin