Validación de la Identificación étnica mediante un sistema de votacion por Mayoría y su ánalisis estadístico en la población de Ecuador

Abstract

As a plurinational and multiethnic State, Ecuador has a diverse population due to its miscegenation. Ethnic groups share physical or genetic traits and characteristics, which allow them to be uniquely identified. The ethnic identification process is complex because it depends on several factors, such as cultural, geographic, economic, and social. A divergence between selfidentification and external identification towards an individual could generate economic and social conflicts in minority Ecuadorian ethnic groups, preventing them from exercising their right to self-determination. This research addresses ethnic identification using a set of facial photographs of the four majority ethnicities in Ecuador: Afro-Ecuadorian, indigenous, European-descended, and mestizos. From these photographs, a majority voting web system was developed using a data structure called a hash table to recover the images and their use in the ethnic voting process. With the data obtained from user interactions, a statistical analysis is performed that measures the divergence between selfperception and external perception of an individual’s ethnicity. The contrast of these two classification processes is summarized in a confusion matrix, where the performance measured with the F1 score is 88.7% in the Afro-Ecuadorian ethnic group, unlike the indigenous ethnic group with 43.7 %, thus showing that people classify certain ethnicities better, based on a facial image. On the other hand, in the ANOVAs, the F test is taken as a reference, resulting in values lower than the significance level established at α = 0.05. These statistical analyses determine that there is no significant difference in invariability in the perception of ethnic groups, which provides a deeper understanding of ethnic selfperception.