The Nonparametric Approach

Abstract

Both the traditional and the more recent parametric techniques in statistics require, as was emphasised in Chapter 2, that a number of assumptions be satisfied. In practice it is true that most of these techniques generate plausible, and even acceptable, solutions even when the assumptions are not wholly satisfied. However, in such cases the claims of optimality cannot be upheld with any conviction, and insufficient evidence is available to justify any great faith in the robustness particularly of the newer geostatistical methods.