(Unless otherwise stated, the copyright of the materials included belong to Jan Woreczko & Wadi.)
Szablon:Jakubowska (2026, MaPS)
Z Wiki.Meteoritica.pl
m |
|||
| Linia 4: | Linia 4: | ||
---- | ---- | ||
'''Abstract:''' The paper presents a modified version of the 4M method, which is the latest method of classifying ordinary chondrites, based on their Mössbauer spectra measured at room temperature. The proposed changes, including the introduction of a new criterion for assessing which group (H, L, or LL) the meteorite being tested belong to, are expected to improve the plausibility of classification by the 4M method. The modification makes use of the Bayesian analysis and the maximum ''a posteriori'' probability. This modified version of the 4M method was tested by attempting to classify 20 samples of ordinary chondrites: 8 of type H, 7 of type L, and 5 of type LL. The results were compared with those obtained by the classical method of ordinary chondrite classification. The vast majority of classification tests performed using the new version of the 4M method were consistent with the classical method for group assignment, except for one L-type sample that was classified differently. It was also shown that the introduction of a new criterion resulted in a significantly better agreement with the established classification than in the case of the level of similarity criterion used in the previous version of the 4M method. | '''Abstract:''' The paper presents a modified version of the 4M method, which is the latest method of classifying ordinary chondrites, based on their Mössbauer spectra measured at room temperature. The proposed changes, including the introduction of a new criterion for assessing which group (H, L, or LL) the meteorite being tested belong to, are expected to improve the plausibility of classification by the 4M method. The modification makes use of the Bayesian analysis and the maximum ''a posteriori'' probability. This modified version of the 4M method was tested by attempting to classify 20 samples of ordinary chondrites: 8 of type H, 7 of type L, and 5 of type LL. The results were compared with those obtained by the classical method of ordinary chondrite classification. The vast majority of classification tests performed using the new version of the 4M method were consistent with the classical method for group assignment, except for one L-type sample that was classified differently. It was also shown that the introduction of a new criterion resulted in a significantly better agreement with the established classification than in the case of the level of similarity criterion used in the previous version of the 4M method. | ||
| + | |||
| + | == Galerie == | ||
| + | |||
| + | Illustrations from the publication (extended version: [http://www.woreczko.pl/meteorites/features/4M-method/4Mb-BASE-vmax_images.htm woreczko.pl]) | ||
| + | <gallery caption="" widths="240px" heights="240px" perrow="3"> | ||
| + | File:Jakubowska_(2026,_MaPS)-4M-BASE_pairs.jpg|'''''FIGURE 1.''' Graphic presentation of the relative areas (expressed as percentages) of four subspectra representing the following types of phases: ol—olivine, pyr—pyroxene, met—metallic phase, and tr—troilite, determined for meteorites constituting the database of the studied and classified ordinary chondrites (symbols: purple—group H, green—group L, yellow—group LL).'' | ||
| + | File:Jakubowska_(2026,_MaPS)-4M-BASE_pie.jpg|'''''FIGURE 2.''' BASE points classification—the 4M-method (pie charts) (validation, training ≡ testing set). The size of each circular graph is proportional to the sum of the levels of similarity (S<sub>clust</sub>). The portions in the pies are proportional to the levels of similarity to individual groups. The accuracy table collects the results of the classification of meteorites from different groups (“true”) related to recognized/predicted groups (“pred”). The correctness of the classification using the 4M method is about 88%.'' | ||
| + | File:Jakubowska_(2026,_MaPS)-4M-BASE_pie-Bayes.jpg|'''''FIGURE 3.''' BASE points classification—Bayes classification rule (pie charts) (validation, training ≡ testing set). The size of each circular graph is proportional to the maximum value of the discriminant function g<sub>k</sub>(x). The portions in the pies are proportional to ''a posteriori'' probabilities for individual groups. The accuracy table collects the results of the classification of meteorites from different groups (“true”) related to recognized/predicted groups (“pred”). The correctness of classification with the revised 4M method is almost 95%.'' | ||
| + | </gallery> | ||
{{Przypisy|ncol=1}} | {{Przypisy|ncol=1}} | ||
| + | |||
| + | == Linki zewnętrzne == | ||
| + | |||
| + | * woreczko.pl – [http://www.woreczko.pl/meteorites/features/4M-method/4Mb-BASE-vmax_images.htm Metoda 4M (Bayes) – 4M-BASE-vmax] | ||
| + | |||
[[Category:Bibliografia (szablony)|{{PAGENAME}}]] | [[Category:Bibliografia (szablony)|{{PAGENAME}}]] | ||
</noinclude> | </noinclude> | ||
Aktualna wersja na dzień 20:22, 20 mar 2026
Jakubowska Martyna, Gałązka-Friedman Jolanta, Woźniak Marek, Szopa Krzysztof, Katarzyna Brzózka, Pospíšilová Barbora, Grabias Agnieszka, (2026), Classification of ordinary chondrites using Mössbauer spectroscopy: Introduction of a new criterion to the 4M method, Meteoritics & Planetary Science, online, 2026, ss. 13 (abstrakt).[1] Plik doi.
Abstract: The paper presents a modified version of the 4M method, which is the latest method of classifying ordinary chondrites, based on their Mössbauer spectra measured at room temperature. The proposed changes, including the introduction of a new criterion for assessing which group (H, L, or LL) the meteorite being tested belong to, are expected to improve the plausibility of classification by the 4M method. The modification makes use of the Bayesian analysis and the maximum a posteriori probability. This modified version of the 4M method was tested by attempting to classify 20 samples of ordinary chondrites: 8 of type H, 7 of type L, and 5 of type LL. The results were compared with those obtained by the classical method of ordinary chondrite classification. The vast majority of classification tests performed using the new version of the 4M method were consistent with the classical method for group assignment, except for one L-type sample that was classified differently. It was also shown that the introduction of a new criterion resulted in a significantly better agreement with the established classification than in the case of the level of similarity criterion used in the previous version of the 4M method.
Galerie
Illustrations from the publication (extended version: woreczko.pl)
FIGURE 1. Graphic presentation of the relative areas (expressed as percentages) of four subspectra representing the following types of phases: ol—olivine, pyr—pyroxene, met—metallic phase, and tr—troilite, determined for meteorites constituting the database of the studied and classified ordinary chondrites (symbols: purple—group H, green—group L, yellow—group LL). |
FIGURE 2. BASE points classification—the 4M-method (pie charts) (validation, training ≡ testing set). The size of each circular graph is proportional to the sum of the levels of similarity (Sclust). The portions in the pies are proportional to the levels of similarity to individual groups. The accuracy table collects the results of the classification of meteorites from different groups (“true”) related to recognized/predicted groups (“pred”). The correctness of the classification using the 4M method is about 88%. |
FIGURE 3. BASE points classification—Bayes classification rule (pie charts) (validation, training ≡ testing set). The size of each circular graph is proportional to the maximum value of the discriminant function gk(x). The portions in the pies are proportional to a posteriori probabilities for individual groups. The accuracy table collects the results of the classification of meteorites from different groups (“true”) related to recognized/predicted groups (“pred”). The correctness of classification with the revised 4M method is almost 95%. |
Przypisy
Linki zewnętrzne
- woreczko.pl – Metoda 4M (Bayes) – 4M-BASE-vmax