Training dataset.
dataset
posted on 2024-07-10, 17:38 authored by Daniel Sanchez-Gomez, Carlos P. Odriozola Lloret, Ana Catarina Sousa, José Ángel Garrido-Cordero, Galo Romero-García, José María Martínez-Blanes, Manel Edo I. Benaiges, Rodrigo Villalobos-García, Victor S. GonçalvesData used to develop the models presented in this study. It contains eight tabs with training and proof-of-concept data, in addition to data and statistics of the reference material, detection limits, imputation values and number of samples of each site considered.
(XLSX)
History
Usage metrics
Categories
Keywords
restrictive class systempersonal adornments dependslargest geochemical datasetfits particularly wellempirical data collectiondifferent validation setsdifferent portuguese sitescompound responses enablingclassification logics givenusing synthetic datacurrent mineralogical characterisation10 samples ).div >< psupervised multiclass frameworkmineral classification workflowmineral classificationreliable characterisationpredictions usingmineral phase15 samples1243 samplesmulticlass modeltwo assemblagestrace provenancesuccessfully assiststraightforward assessmentspecific analysesreusability withoutrespective xpublic repositorypresented approachnecessary resourcesn main limitationmachine learninghandle imbalanceforming mineralsdiscriminative importancecoupling xconsuming routinechemical elementsbenchmarking 13based frameworkallowed us
Licence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC