Repositório Colecção:http://hdl.handle.net/10400.5/702016-08-25T11:27:04Z2016-08-25T11:27:04ZSome misspecfication problems in long-memory time series modelsCrato, Nunohttp://hdl.handle.net/10400.5/119062016-07-29T06:54:03Z1992-12-01T00:00:00ZTítulo: Some misspecfication problems in long-memory time series models
Autor: Crato, Nuno
Resumo: In the middle of this century, the English hydrologist Harold E. Hurst observed remarkably lond cycles on historic records of the Nile river levels. These long non-periodic waves of large amplitude characterized what came to be known as long-memory. In 1968, Mandelbrot and van Ness proposed a new stochastic model in continuous time, the fractional Brownian motion, that was able to explain the long-memory properties of Nile river data in the framework of stationary processes. A discrete-time model that also exhibits long-memory was later advanced by Granger and Joyeux (1980) and, independently, by Hosking (1981). They introduced the fractionally differenced noise, later generalized to fractional autoregressive and moving average processes, FARMA. Since then, interest in long-memory models has increased, and various studies have investigated the long -run dependence of empirical time series, namely in hydrology, climatology, economics and finance.The main purpose of this work is to study forecasting problems that can arise in the presence of long-memory. More specifically, we investigate possible sources and consequences of misspecifications when dealing with long-memory processes. We study in some detail the misinterpretation of a long-memory process as a non-stationary process, leading to an overdifferenced time series and to increased forecasting errors. First, we evaluate the plausibility of this type of misspecification. Second, we charaterize the overdifferenced fractinal noise and show that the observer is likeky to identify it as a low-order ARMA model. Third, we evaluate the increase in forecasting error variance in the cases of both an overdifferenced process with no trend and of an ioverdifferenced process with erroneously identified trend. Finally, we introduce a spectral regression test for stationaritry and derive the periodogram behavior of nonstationary ARIMA processes.
Descrição: Doctor of Philosophy in Mathematics1992-12-01T00:00:00ZO problema da supressão na protecção de informação confidencial: formalizações e algoritmos.Carvalho, Filipa Duarte dehttp://hdl.handle.net/10400.5/117022016-06-27T07:10:27Z2002-09-01T00:00:00ZTítulo: O problema da supressão na protecção de informação confidencial: formalizações e algoritmos.
Autor: Carvalho, Filipa Duarte de
Descrição: Doutoramento em Matemática Aplicada à Economia e à Gestão.2002-09-01T00:00:00ZMethods for routing a vehicle on a bipartite graph at minimum costAlmeida, Maria Tereza Nunes Chaves dehttp://hdl.handle.net/10400.5/112772016-04-03T05:14:25Z1985-03-01T00:00:00ZTítulo: Methods for routing a vehicle on a bipartite graph at minimum cost
Autor: Almeida, Maria Tereza Nunes Chaves de
Resumo: In this thesis heuristic and exact algorithms are developed to
find the minimum cost of routing a vehicle on a bipartite graph
subject to constraints on the number of times each node is to be
visited.
The heuristic method is composed of a construction step which
provides an initial feasible solution, followed by an improvement step
that attempts to derive'a lower cost solution from the initial
one.
The exact method is of branch and bound type, using Lagrangean
relaxation. A specialized method is used to generate .the constraints
to be relaxed, to compute values for their multipliers and to update
the solution, taking advantage of the particular structure of the
problem.
A dynamic programming approach is also investigated, using state
space relaxation and a penalty method to improve the bound. A
comparative study of the computational results obtained on a set of
test problems was not favourable to this approach.
Computational experience is reported for all methods on euclidean
and randomly generated cost matrix problems.
Descrição: Doutoramento em Matemática1985-03-01T00:00:00ZThe use of semi-parametric methods in achieving robust inferencePassos, José Manuel de Matoshttp://hdl.handle.net/10400.5/112722016-04-03T05:14:13Z1996-05-01T00:00:00ZTítulo: The use of semi-parametric methods in achieving robust inference
Autor: Passos, José Manuel de Matos
Resumo: This thesis focuses on some topics in semi-parametric econometrics, particularly
the use of semi-parametric methods of estimation to obtain robust inference.
Chapter two proposes a study of the finite-sample performance of the heteroskedastic
and autocorrelation consistent covariance matrix estimators (HAC). This performance
is accessed through the bias of the first moment of HAC type estimators and
the quality of the asymptotic normal approximation to the exact finite-sample distributions
of HAC type Wald statistics of scalar linear hypothesis.
In Chapter three, the use of the non-overlapping deleted-l jackknife is used to propose
a new approach to estimate the covariance matrix of the least square estimator
in a linear regression model. This estimator is robust to the presence of heteroskedastldty
and autocorrelation in the errors.
Chapter four deals with improved estimation of regression coefficients through an
alternative and efficient method of estimation regression models under heteroskedasticity
of tmknown form. Kernel and average derivative estimation are used to estimate
the conditional variance of the response variable where this conditional variance is
assumed to be in an index form.
Chapter five is concerned with the estimation of duration models under unobserved
heterogeneity. This is a typical problem in mlcroer.onometrics and is in general due
to differences among individuals. It is suggested a method of estimation based on a
roughness penalty approach.
Descrição: Doutoramento em Matemática1996-05-01T00:00:00Z