IdentificaÃÃo de Pontos Influentes em uma Amostra AleatÃria de PrÃ-formas da DistribuiÃÃo Bingham Complexa (DistÃncia de Cook e MÃtodos de Bootstrap)

AUTOR(ES)

Olga PatrÃcia Reyes FlÃrez

DATA DE PUBLICAÇÃO

2009

RESUMO

The objective of this dissertation is the evaluation and application of influence analysis methods in statistical shape analysis. Through the model of deleting cases (CDM) is obtained a measure of the Cook distance when the data set has the complex Bingham distribution. Through simulations of Monte Carlo and the bootstrap method the estimation of region of confidence for the distance of Cook is gotten. By the Monte Carlo simulation and the Bootstrap methods the confidence interval is estimated for the Cook distance in different data sets with complex Bingham distribution. Beyond that, it is shown that other influence analysis methods can work on shape analysis. The efficacy of the Cook distance in the presented methods is evaluated.

ASSUNTO(S)

bingham complexa cook distance distÃncia de cook bingham complex anÃlise de formas estatistica bootstrap shape analysis. bootstrap

Documentos Relacionados

• AdiÃÃo seqÃencial aleatÃria de elementos escalantes em d dimensÃes
• AnÃlise da InfluÃncia do Fator DistribuiÃÃo Espacial dos Dados no Desempenho de MÃtodos de Acesso Multidimensionais
• A distribuiÃÃo da competÃncia legislativa tributÃria no Brasil e a obrigatoriedade do seu exercÃcio
• Ondas de spin em sistemas magnÃticos bidimensionais com interaÃÃo de troca aleatÃria.
• Sistemas Java de educaÃÃo à distÃncia na internet