I.6.5.4: Squared Distance Scaling

Suppose we want to minimize with squared Euclidean distance. An augmentation algorithm for this problem, modestly called ELEGANT, was designed by De Leeuw [1975]. That paper was never published and the manuscript is probably lost, but the algorithm was described, discussed, and applied by both Takane [1977] and Browne [1987]

We augment to where we require while the other are free. Define and assume that is column-centered, i.e. is doubly-centered. The augmentation works because Also where Thus we can minimize the augmented loss function over for fixed by minimizing . This means computing the largest eigenvalues and corresponding eigenvectors of .

Minimizing the augmented loss over the for fixed is This is enough information to get the ALS algorithm going. The "elegance" so far is reducing a problem involving multivariate quartics to a sequence of eigenvalue problems. It is distinctly unelegant, however, that the computations need four-dimensional arrays. But it turns out these can easily be gotten rid of. We use

where and . Thus we find by computing eigenvalues and eigenvectors of and no intermediate computation or storage of the is required.