Next: Robust Determination of
Up: 5: Goodness of
Previous: Robust Estimation---Motivation---II
- There is thus a great deal of interest in the study of so called
robust techniques which are insensitive to errors in our
knowledge about the underlying distribution.
- Statisticians are very interested in such techniques but (apart
from non-parametric methods we have already discussed) there has been
little use/experience of them in the type of hard problems that
physicists are want to find.
- Further, the derivation of robust methods is rather intuitive and
not as mathematically precise as those of classical statistics.
Geoffrey Fox, Northeast Parallel Architectures Center at Syracuse University, gcf@npac.syr.edu