4Mb method (Bayes) – form (script)  « «

4M-BASE – all versions  (→ illustrations)

Using the form (script) below, you can calculate the Mahalanobis distances (dM) and the MAP (maximum a posteriori) for an unknown meteorite using the 4M method.

gj(xi) = –0.5 * dMj * dMj – 0.5 * ln(detSj) + ln(pj)

So make a classification – what type is it – H, L or LL! (For more details on the interpretation of the results MAP (maximum a posteriori) see → Footnotes)

Enter the percentage values of the spectral areas obtained from Mössbauer spectroscopy in the text boxes below:

  • ol – value for olivine;

  • pyr – value for pyroxene;

  • met – cumulative value for the metallic phase (kamacite+taenite+tetrataenite);

  • tr – cumulative value for the sulfide phase (troilite+pyrrhotite).

 

ol (olivine) pyr (pyroxene) met (metal) tr (troilite) (optional sample name)

 

(The script is based on the R statistics package installed on the server)

Meteorite is assigned to the type for which the MAP (maximum a posteriori) value is the biggest!

The results of the next calculations will be displayed here...

*————

Footnotes

A t t e n t i o n [-1]:  The result is inconclusive (niejednoznaczny)! Uwaga! Minimalna odl. Mahalanobisa dM wskazuje na inny typ niż maksymalna wartość kryterium g() i/lub mała różnica wartości g() (<10%) dla dwóch typów ---- Difference in g() too small (<10%), the results lie „on the border” between two different types ▶→ perhaps: it is an intermediate type chondrite: H/L, L/LL or breccia?!

A t t e n t i o n [-2]:  The result is unreliable (niepewny)! Duża wartość minimalnej odl. Mahalanobisa (dM>3σ) ▶→ perhaps: the specimen is heavily weathered / breccia / chondrite unequilibrated / incorrectly fitting results?!

A t t e n t i o n [-3]:  Unbelievable result (niewiarygodny)! The maximum ---- is too low (<5%) ▶→ perhaps: value error / bad, incorrectly prepared sample / incorrect fit of the results / it's not chondrite...?!

A t t e n t i o n [-4]:  Incorrect data!?! Not everything is ordinary chondrite (<1%)

A t t e n t i o n [∞]:  BINGO! You've come in the center of cluster – such a low does not guarantee certainty

Acknowledgments

We thank the author of the scripts, Janusz "Johnson" Lemański. Without his knowledge and commitment, this website would not be created. We appreciate his patience and understanding for our whining, vague expectations and "lame" problems ;-)