SUPPORT VECTOR MACHINE-BASED CLASSIFICATION OF PALEOZOIC GYMNOSPERM WOODS
Clasificación de maderas de gimnospermas paleozoicas basada en máquinas de soporte vectorial
DOI:
https://doi.org/10.5710/PEAPA.25.06.2026.562Keywords:
Supervised learning, Wood anatomy, Taxonomy, Computational paleontology, PaleobotanyAbstract
Understanding the taxonomic placement of Paleozoic gymnosperm woods remains a complex task largely due to the frequent absence of attached reproductive or foliar structures. Machine-assisted analyses based on wood anatomy can support this task by suggesting plausible affinities between fossil genera and taxonomic orders for paleobotanists to further explore. In this context, we developed a support vector classifier employing a soft-margin, which achieved a maximum estimated accuracy of 71.4% and a macro-averaged F1-Score of 61% across four evaluated regularization values, representing an improvement over previously proposed models. The null hypothesis of the permutation test by label shuffling was rejected (p < 0.001), indicating that the model effectively learned the underlying structure linking anatomical features to taxonomic orders. To evaluate its applicability, six genera from western Gondwana were selected as input to test the classifier, with predicted affinities either aligning with previous literature or offering new perspectives for future paleobotanical research. The model proposed is data-driven and capable of learning from patterns in anatomical characters, operating through statistical generalization across examples. While the model’s main limitations are rooted in biological and taphonomical constraints, its overall performance demonstrates the usefulness of machine learning approaches as an additional aid to fossil wood taxonomy.
References
Awad, M., & Khanna, R. (2015). Support vector machines for classification. In Awad, M., & Khanna, R. (Eds.), Efficient learning machines: Theories, concepts, and applications for engineers and system designers (pp. 39–66). Apress. https://doi.org/10.1007/978-1-4302-5990-9_3
Burges, C. J. (1998). A tutorial on support vector machines for pattern recognition. Data mining and knowledge discovery, 2(2), 121–167. https://doi.org/10.1023/A:1009715923555
Caldas, E. B., Mussa, D., Lima Filho, F. P., & Rösler, O. (1989). Nota sobre a ocorrência de uma floresta petrificada de idade permiana em Teresina, Piauí. Boletim IG-USP. Publicação Especial, 7, 69–87. https://doi.org/10.11606/issn.2317-8078.v0i7p69-87 [in Portuguese]
Cervantes, J., Garcia-Lamont, F., Rodríguez-Mazahua, L., & Lopez, A. (2020). A comprehensive survey on support vector machine classification: Applications, challenges and trends. Neurocomputing, 408, 189–215. https://doi.org/10.1016/j.neucom.2019.10.118
Césari, S., Archangelsky, S., & Vega, J. (2005). Anatomy of a new probable pteridosperm stem from the Late Carboniferous of Argentina. Revista del Museo Argentino de Ciencias Naturales, Nueva Serie, 7(1), 7–15.
Chang, C. C., & Lin, C. J. (2011). LIBSVM: a library for support vector machines. ACM transactions on intelligent systems and technology (TIST), 2(3), 1–27. https://doi.org/10.1145/1961189.1961199
Coimbra, A. M., & Mussa, D. (1984). Associação lignitafoflorística da Formação Pedra-de-Fogo,(Arenito Cacunda), Bacia do Maranhão-Piauí, Brasil. Anais do XXXIII Congresso Brasileiro de Geologia (v 2, pp. 591–605). Rio de Janeiro. [in Portuguese]
Conceição, D. M., Crisafulli, A., Iannuzzi, R., Neregato, R., Cisneros, J. C., & Andrade, L. S. (2020a). New petrified gymnosperms from the Permian of Maranhão (Pedra de Fogo formation), Brazil: Novaiorquepitys and Yvyrapitys. Review of Palaeobotany and Palynology, 276, 104177. https://doi.org/10.1016/j.revpalbo.2020.104177
Conceição, D. M., Iannuzzi, R., Andrade, L. S., Esperança Júnior, M. G. F., Cisneros, J. C., & Siegloch, A. M. (2020b). First report of Cordaixylon Grand'Eury in the Permian of South America, Parnaíba Basin, Brazil. Journal of South American Earth Sciences, 101, 102620. https://doi.org/10.1016/j.jsames.2020.102620
Conceição, D. M., Esperança Júnior, M. G. F., Iannuzzi, R., & Cisneros, J. C. (2022). Two new petrified gymnosperms with solenoid piths from the Pedra de Fogo Formation, Permian of Maranhão, Brazil. Review of Palaeobotany and Palynology, 299, 104622. https://doi.org/10.1016/j.revpalbo.2022.104622
Conceição, D. M., Esperança Júnior, M. G. F., Iannuzzi, R., Recamonde-Mendoza, M., & Malta, G. B. B. O. (2023). PaleoWood: A machine learning approach for determining the affinity of Paleozoic gymnosperm woods. Journal of South American Earth Sciences, 121, 104125. https://doi.org/10.1016/j.jsames.2022.104125
Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20, 273–297. https://doi.org/10.1007/BF00994018
Crisafulli, A. (1998). Leños gimnospérmicos de la Formación Melo (Pérmico Inferior), Uruguay. Parte III. Paulistoxylon, Austroscleromedulloxylon y Piracicaboxylon. Ameghiniana, 35(2), 217-225.
Crisafulli, A. (2001). Leños pérmicos de la Formación Yaguarí, República Oriental del Uruguay. (2014). Ameghiniana, 38(1), 61–72.
Crisafulli, A., & Lutz, A. (2007). Eoguptioxylon antiqua gen. et sp. nov.(Pteridospermae) del Pérmico Superior (Formación la Antigua), provincia de la Rioja, Argentina. Ameghiniana, 44(1), 197–204.
Cristianini, N., & Ricci, E. (2008). Support vector machines. In Kao, M.Y. (Ed.), Encyclopedia of algorithms (pp. 928–932). Springer. https://doi.org/10.1007/978-0-387-30162-4_415
Decombeix, A. L., Meyer-Berthaud, B., & Galtier, J. (2011). Transitional changes in arborescent lignophytes at the Devonian–Carboniferous boundary. Journal of the Geological Society, 168(2), 547–557. https://doi.org/10.1144/0016-76492010-074
Decombeix, A. L., Galtier, J., McLoughlin, S., Meyer-Berthaud, B., Webb, G. E., & Blake, P. R. (2019). Early Carboniferous lignophyte tree diversity in Australia: Woods from the Drummond and Yarrol basins, Queensland. Review of Palaeobotany and Palynology, 263, 47-64. https://doi.org/10.1016/j.revpalbo.2019.01.009
Dunn, M. T. (2006). A review of permineralized seed fern stems of the Upper Paleozoic. The Journal of the Torrey Botanical Society, 133(1), 20–32. https://doi.org/10.3159/1095-5674(2006)133[20:AROPSF]2.0.CO;2
Ferguson, D. K. (2005). Plant taphonomy: ruminations on the past, the present, and the future. Palaios, 20(5), 418-428. https://doi.org/10.2110/palo.2005.P05-25p
Galtier, J., & Meyer-Berthaud, B. (2006). The diversification of early arborescent seed ferns. The Journal of the Torrey Botanical Society, 133(1), 7–19. https://doi.org/10.3159/1095-5674(2006)133[7:TDOEAS]2.0.CO;2
Greenwood, D.R. (1991). The taphonomy of plant macrofossils. In S.K. Donovan (Ed.), The Processes of Fossilization (pp. 141–169), Columbia University Press.
Guerra-Sommer, M. (1977). Damudoxylon (Maheshwari) Maheshwari, 1972, um gênero ocorrente no Gondwana do Brasil. Pesquisas em Geociências, 7(7), 131–144. https://doi.org/10.22456/1807-9806.21824 [in Portuguese]
Hearst, M. A., Dumais, S. T., Osuna, E., Platt, J., & Scholkopf, B. (1998). Support vector machines. IEEE Intelligent Systems and their applications, 13(4), 18–28. https://doi.org/10.1109/5254.708428
Hernandez-Castillo, G. R., Rothwell, G. W., & Mapes, G. (2001). Thucydiaceae fam. nov., with a review and reevaluation of Paleozoic walchian conifers. International Journal of Plant Sciences, 162(5), 1155–1185. https://doi.org/10.1086/321920
Iannuzzi, R., & Vieira, C. E. L. (2005). Paleobotânica. Editora da UFRGS. [in Portuguese]
James, G., Witten, D., Hastie, T., & Tibshirani, R. (2013). An introduction to statistical learning. Springer.
Kurzawe, F., Iannuzzi, R., Merlotti, S., Rößler, R., & Noll, R. (2013). New gymnospermous woods from the Permian of the Parnaíba Basin, northeastern Brazil, part I: Ductoabietoxylon, Scleroabietoxylon and Parnaiboxylon. Review of Palaeobotany and Palynology, 195, 37–49. https://doi.org/10.1016/j.revpalbo.2012.12.004
Lacey, W. S. (1953). Scottish Lower Carboniferous Plants: Eristophyton waltoni sp. nov. and Endoxylon zonatum (Kidston) Scott from Dunbartonshire. Annals of Botany, 17(4), 579–597. https://doi.org/10.1093/oxfordjournals.aob.a083371
Liu, Y., Torres, L. N., Wang, B., Pei, W., Na, Y., Song, Q., & Shi, X. (2025). Artificial Intelligence in Paleobotany and Palynology. Geological Journal. https://doi.org/10.1002/gj.70007
Lorena, A. C., & De Carvalho, A. C. (2007). Uma introdução às support vector machines. Revista de Informática Teórica e Aplicada, 14(2), 43–67. https://doi.org/10.22456/2175-2745.5690 [in Portuguese]
Ma, Y., & Guo, G. (2014). Support vector machines applications. Springer. https://doi.org/10.1007/978-3-319-02300-7
Mammone, A., Turchi, M., & Cristianini, N. (2009). Support vector machines. Wiley Interdisciplinary Reviews: Computational Statistics, 1(3), 283–289. https://doi.org/10.1002/wics.49
Merlotti, S. (2002). Dois novos taxa lenhosos da Formação Serra Alta (Permiano Superior, Bacia do Paraná), estado do Rio Grande do Sul, Brasil. Revista Geociências – UNG-Ser – ISSN 1981-741X (1981-7428), 7(6), 5–14. [in Portuguese]
Merlotti, S. (2009). Reavaliação taxonômica de lenhos das formações Irati e Serra Alta, Permiano da Bacia do Paraná, Brasil. Pesquisas em Geociências, 36(1), 11–21. https://doi.org/10.22456/1807-9806.17871 [in Portuguese]
Meyer, D., Dimitriadou, E., Hornik, K., Weingessel, A., & Leisch, F. (2024). Package ‘e1071’. https://doi.org/10.32614/CRAN.package.e1071
Moguerza, J. M., & Muñoz, A. (2006). Support vector machines with applications. Statistical Science, 21(3), 322–336. https://doi.org/10.1214/088342306000000493
Mussa, D. (1978). Brasilestiloxylon e Solenobrasilioxylon, dois novos gêneros gondwânicos na formação Irati, Boletim IG-USP, 9, 118–127. https://doi.org/10.11606/issn.2316-8978.v9i0p118-127 [in Portuguese]
Mussa, D. (1986a). Eustelos Gondwânicos de medulas diafragmadas e a sua posição estratigráfica. Boletim IG-USP. Série Científica, 17, 11–24. https://doi.org/10.11606/issn.2316-8986.v17i0p11-24 [in Portuguese]
Mussa, D. (1986b). As formas gondwânicas do Grupo Solenóide e sua distribuição estratigráfica. Anais da Academia Brasileira de Ciências, 58(1), 61–88. [in Portuguese]
Ojala, M., & Garriga, G. C. (2010). Permutation tests for studying classifier performance. Journal of machine learning research, 11(6).
Pigg, K. B., Taylor, T. N., & Stockey, R. A. (1987). Paleozoic seed ferns: Heterangium kentuckyensis sp. nov., from the Upper Carboniferous of North America. American Journal of Botany, 74(8), 1184–1204. https://doi.org/10.1002/j.1537-2197.1987.tb08733.x
Platt, J. (1998). Sequential minimal optimization: A fast algorithm for training support vector machines. Microsoft Research.
Read, C. B. (1936). The flora of the New Albany shale: Part 1, Diichnia kentuckiensis, a new representative of the Calamopityeae. U.S. Geological Survey professional paper, 185–H, 149–161.
Rößler, R., Philippe, M., van Konijnenburg-van Cittert, J. H., McLoughlin, S., Sakala, J., Zijlstra, G., Annamraju, R., Ash, S., Baas, P., Bamford, M., Bateman, R., Booi, M., Boonchai, N., Brea, M., Crisafulli, A., Decombeix, A.L., Dolezych, M., Dutra, T., Esteban, L.G., ... Zhang, W. (2014). Which name (s) should be used for Araucaria‐like fossil wood?—Results of a poll. Taxon, 63(1), 177–184. https://doi.org/10.12705/631.7
Shi-Jun, W., Hilton, J., Baolin, T., & Galtier, J. (2003). Cordaitalean seed plants from the Early Permian of North China. I. Delimitation and reconstruction of the Shanxioxylon sinense plant. International Journal of Plant Sciences, 164(1), 89–112. https://doi.org/10.1086/344558
Simões, M. G., Rodrigues, S. C., Bertoni-Machado, M., & Holz, M. (2010). Tafonomia: processos e ambientes de fossilização. In Carvalho, I. S. (Ed.), Paleontologia (pp. 19–51), Interciência. [in Portuguese]
Sokolova, M., & Lapalme, G. (2009). A systematic analysis of performance measures for classification tasks. Information processing & management, 45(4), 427–437. https://doi.org/10.1016/j.ipm.2009.03.002
Spicer, R. A. (1991). Plant taphonomic processes. In P. A. Allison & D. Briggs (Eds.), Taphonomy: Releasing the data locked in the fossil record (pp. 71–113), Springer.
Spiekermann, R., Jasper, A., Siegloch, A. M., Guerra-Sommer, M., & Uhl, D. (2021). Not a lycopsid but a cycad-like plant: Iratinia australis gen. nov. et sp. nov. from the Irati Formation, Kungurian of the Paraná Basin, Brazil. Review of palaeobotany and palynology, 289, 104415. https://doi.org/10.1016/j.revpalbo.2021.104415
Taylor, E. L., Taylor, T. N., & Krings, M. (2009). Paleobotany: the biology and evolution of fossil plants. Academic Press.
Trivett, M. L., & Rothwell, G. W. (1985). Morphology, systematics, and paleoecology of Paleozoic fossil plants: Mesoxylon priapi, sp. nov.(Cordaitales). Systematic Botany, 10, 205–223. https://doi.org/10.2307/2418347
Tsochantaridis, I., Joachims, T., Hofmann, T., & Altun, Y. (2005). Large margin methods for structured and interdependent output variables. Journal of machine learning research, 6, 1453–1484.
Wan, M., Shi, G. R., Luo, M., Lee, S., & Wang, J. (2020). First record of a petrified gymnospermous wood from the Kungurian (late Early Permian) of the southern Sydney Basin, southeastern Australia, and its paleoclimatic implications. Review of Palaeobotany and Palynology, 276, 104202. https://doi.org/10.1016/j.revpalbo.2020.104202
Wang, K., Huang, X., Yang, W., Wang, J., & Wan, M. (2023). A new gymnospermous stem from the Moscovian (Carboniferous) of North China, and its palaeoecological significance for the Cathaysian Flora at the early evolutionary stage. Review of Palaeobotany and Palynology, 311, 104858. https://doi.org/10.1016/j.revpalbo.2023.104858
Wang, K., Jia, G., Dong, L., Wang, J., Wang, S., Wang, J., & Wan, M. (2025). Shanxioxylon yangquanense sp. nov., a new Kasimovian cordaitalean axis from the Benxi Formation (Pennsylvanian, Carboniferous) of Yangquan City, Shanxi Province, North China. Review of Palaeobotany and Palynology, 335, 105287. https://doi.org/10.1016/j.revpalbo.2025.105287
Yu, C., Qin, F., Watanabe, A., Yao, W., Li, Y., Qin, Z., Liu, Y., Wang, H., Jiangzuo, Q., Hsiang, A.Y., Ma, C., Rayfield, E., Benton, M.J., & Xu, X. (2024). Artificial intelligence in paleontology. Earth-Science Reviews, 104765. https://doi.org/10.1016/j.earscirev.2024.104765
Additional Files
Published
Issue
Section
License
Copyright (c) 2026 Mário G. F. Esperança Júnior, Roberto Iannuzzi, Domingas Maria da Conceição
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.
Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Atribución/Reconocimiento 4.0 Internacional that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
0.299 Q3 | H Index 12
CiteScore 2.2
