Probabilistic Analysis of Conic Condition Numbers: II. The Real Case
Our final goal can be succinctly stated: we want to extend the main result of the preceding chapter to problems with real data. Again this boils down to providing bounds on the volume of tubes. However, the technical development is quite different. One of the key constituents of the proof of the main result of Chap. 20 is false over the reals. For this reason, a more sophisticated line of argument involving integrals of curvature and the kinematic formula is required.