English
 
Privacy Policy Disclaimer
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT
Add to Basket
  Local TagsRelease HistoryDetailsSummary

Released
Conference Paper

RMS - Rather Meaningless Simplification?

Authors
/persons/resource/agrayver

Grayver,  Alexander
2.2 Geophysical Deep Sounding, 2.0 Physics of the Earth, Departments, GFZ Publication Database, Deutsches GeoForschungsZentrum;
25. Kolloquium, 2013, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung, External Organizations;

/persons/resource/ktietze

Tietze,  Kristina
2.2 Geophysical Deep Sounding, 2.0 Physics of the Earth, Departments, GFZ Publication Database, Deutsches GeoForschungsZentrum;
25. Kolloquium, 2013, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung, External Organizations;

/persons/resource/oritter

Ritter,  Oliver
2.2 Geophysical Deep Sounding, 2.0 Physics of the Earth, Departments, GFZ Publication Database, Deutsches GeoForschungsZentrum;
25. Kolloquium, 2013, Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung, External Organizations;

External Ressource
Fulltext (public)

EMTF_2013_31-35.pdf
(Publisher version), 2MB

Supplementary Material (public)
There is no public supplementary material available
Citation

Grayver, A., Tietze, K., Ritter, O. (2013): RMS - Rather Meaningless Simplification?, 25. Schmucker-Weidelt-Kolloquium für Elektromagnetische Tiefenforschung (Kirchhundem Rahrbach, Germany 2013), 31-35.


Cite as: https://gfzpublic.gfz-potsdam.de/pubman/item/item_248022
Abstract
For a long time the root mean square (RMS) error has been used in the EM community: - to characterize data fit for a particular model; - as a criterion to compare several models obtained from inversion. The RMS error appears to be a natural choice since we usually tackle inverse problems in a least-squares sense. Over the years, RMS became a customary criterion and gained ultimate significance. However, on the hunt for low RMS values, one often needs to introduce subjectivity by arbitrarily adjusting error floors or masking “bad” data without referring to the assumptions behind RMS. In this contribution, we revisit basic assumptions behind RMS, demonstrate its deficiency and propose alternative ways, which may provide more insight into our data and allow a more comprehensive assessment of the quality of the modelling result/resistivity model.