| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 281.92 KB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Os preços diários dos ativos financeiros exibem comportamentos e características estatísticas comuns. Torna-se então interessante identificar semelhanças entre as séries financeiras, podendo depois classificá-las e agrupá-las. O uso de técnicas de análise multivariada, nomeadamente de clustering, mostram-se de extrema importância para a atividade financeira. Neste trabalho compara-se o desempenho de medidas de distância baseadas nos coeficientes estimados da função de autocorrelação do valor absoluto dos retornos com o quadrado dos retornos. Uma aplicação empírica destas técnicas ao índice FTSE100 é também apresentada, permitindo a comparação do comportamento da dependência a curto prazo com a dependência a longo prazo.
The daily prices of financial assets exhibit some behavior and statistical properties in common. It becomes interesting to identify the similarities between financial time series, and to be abble to classify and group them. The use of tehcniques of multivariate analysis, namely clustering, is of extreme importance to the financial activity. In this work the performance of a distance measure based on the estimated autocorrelation coefficients of the absolute value of the returns is compared with the square of the returns. An empirical application of these techniques based on the FTSE100 index is also presented, allowing a comparision between the behavior of the short-term and the long-term dependence.
The daily prices of financial assets exhibit some behavior and statistical properties in common. It becomes interesting to identify the similarities between financial time series, and to be abble to classify and group them. The use of tehcniques of multivariate analysis, namely clustering, is of extreme importance to the financial activity. In this work the performance of a distance measure based on the estimated autocorrelation coefficients of the absolute value of the returns is compared with the square of the returns. An empirical application of these techniques based on the FTSE100 index is also presented, allowing a comparision between the behavior of the short-term and the long-term dependence.
Descrição
Mestrado em Matemática Financeira
Palavras-chave
clustering séries financeiras medidas de distância factos estilizados modelos heterocedásticos financial series distance measure stylized facts heteroscedastic models
Contexto Educativo
Citação
Parrinha, Filipa Martins (2012). "Clusters de séries financeiras". Dissertação de Mestrado, Universidade Técnica de Lisboa. Instituto Superior de Economia e Gestão.
Editora
Instituto Superior de Economia e Gestão
