Publicação

Weighted coefficients to measure agreement among several sets of ranks emphasizing top and bottom ranks at the same time

Ver documento

Detalhes bibliográficos
Resumo:In this paper, two new weighted coefficients of agreement to measure the concordance among several (more than two) sets of ranks, putting more weight in the lower and upper ranks simultaneously, are presented. These new coefficients, the signed Klotz and the signed Mood, generalize the correspondent rank-order coefficients to measure the agreement between two sets of ranks previously proposed by the authors [1]. Under the null hypothesis of no agreement or no association among the rankings, the asymptotic distribution of these new coefficients was derived. To illustrate the worth of these measures, an example is presented to compare them with the Kendall’s coefficient and the van der Waerden correlation coefficient.
Autores principais:Aleixo, Sandra
Outros Autores:Teles, Júlia
Assunto:Rank-order correlation Weighted coefficients · Concordance measures Signed Klotz coefficient · Signed Mood coeficiente van der Waerden coefficient
Ano:2019
País:Portugal
Tipo de documento:documento de conferência
Tipo de acesso:acesso restrito
Instituição associada:Instituto Politécnico de Lisboa
Idioma:inglês
Origem:Repositório Científico do Instituto Politécnico de Lisboa
Descrição
Resumo:In this paper, two new weighted coefficients of agreement to measure the concordance among several (more than two) sets of ranks, putting more weight in the lower and upper ranks simultaneously, are presented. These new coefficients, the signed Klotz and the signed Mood, generalize the correspondent rank-order coefficients to measure the agreement between two sets of ranks previously proposed by the authors [1]. Under the null hypothesis of no agreement or no association among the rankings, the asymptotic distribution of these new coefficients was derived. To illustrate the worth of these measures, an example is presented to compare them with the Kendall’s coefficient and the van der Waerden correlation coefficient.