Análisis Estadístico Multivariante de los Principales Componentes Químicos sobre Hachas Monedas de Cobre para Determinar Diferentes Grupos de Acuerdo a su Composición

Abstract

The main objective of this research was to perform a multivariate statistical analysis of the main chemical components found on copper coin axes in order to determine different groups according to their chemical composition. The application of such techniques has been done within an archaeological context, the analyzed axes belong to a private collection of EcuaCorriente S. A. This research was carried out under the parameters of a quantitative experimental research. To initiate the study it was necessary to carry out a restoration process to the archaeological pieces prior to the measurement of the chemical composition of the copper coin axes with the Thermo Niton machine, all this procedure was carried out in the archaeological laboratory of ECSA located in the Tundayme Parish, Canton El Pangui, Province of Zamora Chinchipe. In the study of the multiple linear regression of each variable on the remaining ones, five models were chosen with the purpose of making predictions as their squared determination coefficients are between 96% and 99%. When applying the cluster analysis, the presence of 6 groups was detected, each of which is constituted by coin axes of similar chemical characteristics. With the application of the factorial analysis, the presence of latent variables was detected, which suggest that 60% of the coin axes are homogeneously distributed and therefore they are formed fundamentally by the alloy of copper and arsenic. Therefore, it is concluded that the multivariate statistical techniques applied in the investigation were of great help to determine the groups of axes formed according to their chemical composition. Finally, it is recommended to promote this type of studies that not only promote the application of statistical culture, but also strengthen and contribute to the history and identity of Ecuador.