Transformations and Algorithms for Least Sum of Squares Hypersphere Fitting
Michael Burr, Alan C. Cheng, Ryan G. Coleman, Diane L. Souvaine · Canadian Conference on Computational Geometry · 2004
A problem from shape fitting is finding the sphere which has the least sum of squares fit to a set of points. Most current algorithms for this use an iterative process, with randomized restarting. We present, instead, two geometric transformations, and algorithms derived using them, which present alternate methods of solving the problem. We discuss running of these algorithms in three dimensions, and note the algorithms have favorable time complexities in high dimension. Finally, we present directions for future work.