A New Model Of Trend Inflation

This paper introduces a new model of trend (or underlying) inflation. In contrast to many earlier approaches, which allow for trend inflation to evolve according to a random walk, ours is a bounded model which ensures that trend inflation is constrained to lie in an interval. The bounds of this interval can either be fixed or estimated from the data. Our model also allows for a time-varying degree of persistence in the transitory component of inflation. The bounds placed on trend inflation mean that standard econometric methods for estimating linear Gaussian state space models cannot be used and we develop a posterior simulation algorithm for estimating the bounded trend inflation model. In an empirical exercise with CPI inflation we find the model to work well, yielding more sensible measures of trend inflation and forecasting better than popular alternatives such as the unobserved components stochastic volatility model.

Large Time-Varying Parameter VARs

In this paper we develop methods for estimation and forecasting in large timevarying parameter vector autoregressive models (TVP-VARs). To overcome computational constraints with likelihood-based estimation of large systems, we rely on Kalman filter estimation with forgetting factors. We also draw on ideas from the dynamic model averaging literature and extend the TVP-VAR so that its dimension can change over time. A final extension lies in the development of a new method for estimating, in a time-varying manner, the parameter(s) of the shrinkage priors commonly-used with large VARs. These extensions are operationalized through the use of forgetting factor methods and are, thus, computationally simple. An empirical application involving forecasting inflation, real output, and interest rates demonstrates the feasibility and usefulness of our approach.

Model Switching and Model Averaging in Time- Varying Parameter Regression Models

This paper investigates the usefulness of switching Gaussian state space models as a tool for implementing dynamic model selecting (DMS) or averaging (DMA) in time-varying parameter regression models. DMS methods allow for model switching, where a different model can be chosen at each point in time. Thus, they allow for the explanatory variables in the time-varying parameter regression model to change over time. DMA will carry out model averaging in a time-varying manner. We compare our exact approach to DMA/DMS to a popular existing procedure which relies on the use of forgetting factor approximations. In an application, we use DMS to select different predictors in an in ation forecasting application. We also compare different ways of implementing DMA/DMS and investigate whether they lead to similar results.

A low dimensional Kalman filter for systems with lagged observables

This note describes how the Kalman filter can be modified to allow for thevector of observables to be a function of lagged variables without increasing the dimensionof the state vector in the filter. This is useful in applications where it is desirable to keepthe dimension of the state vector low. The modified filter and accompanying code (whichnests the standard filter) can be used to compute (i) the steady state Kalman filter (ii) thelog likelihood of a parameterized state space model conditional on a history of observables(iii) a smoothed estimate of latent state variables and (iv) a draw from the distribution oflatent states conditional on a history of observables.

State-space approach to evaluate the relation between soil physical and chemical properties

The state-space approach is used to evaluate the relation between soil physical and chemical properties in an area cultivated with sugarcane. The experiment was carried out on a Rhodic Kandiudalf in Piracicaba, State of São Paulo, Brazil. Sugarcane was planted on an area of 0.21 ha i.e., in 15 rows 100 m long, spaced 1.4 m. Soil water content, soil organic matter, clay content and aggregate stability were sampled along a transect of 84 points, meter by meter. The state-space approach is used to evaluate how the soil water content is affected by itself and by soil organic matter, clay content, and aggregate stability of neighboring locations, in different combinations, aiming to contribute to a better understanding of the relation among these variables in the soil. Results show that soil water contents were successfully estimated by this approach. Best performances were found when the estimate of soil water content at locations i was related to soil water content, clay content and aggregate stability at locations i-1. Results also indicate that this state-space model using all series describes the soil water content better than any equivalent multiple regression equation.

Potencialidad de la modelización state-space y el filtro de Kalman para el anàlisis regional : una aplicación para el índice de actividad industrial

El déficit existente a nuestro país con respecto a la disponibilidad de indicadores cuantitativos con los que llevar a término un análisis coyuntural de la actividad industrial regional ha abierto un debate centrado en el estudio de cuál es la metodología más adecuada para elaborar indicadores de estas características. Dentro de este marco, en este trabajo se presentan las principales conclusiones obtenidas en anteriores estudios (Clar, et. al., 1997a, 1997b y 1998) sobre la idoneidad de extender las metodologías que actualmente se están aplicando a las regiones españolas para elaborar indicadores de la actividad industrial mediante métodos indirectos. Estas conclusiones llevan a plantear una estrategia distinta a las que actualmente se vienen aplicando. En concreto, se propone (siguiendo a Israilevich y Kuttner, 1993) un modelo de variables latentes para estimar el indicador de la producción industrial regional. Este tipo de modelo puede especificarse en términos de un modelo statespace y estimarse mediante el filtro de Kalman. Para validar la metodología propuesta se estiman unos indicadores de acuerdo con ella para tres de las cuatro regiones españolas que disponen d¿un Índice de Producción Industrial (IPI) elaborado mediante el método directo (Andalucía, Asturias y el País Vasco) y se comparan con los IPIs publicados (oficiales). Los resultados obtenidos muestran el buen comportamiento de l¿estrategia propuesta, abriendo así una línea de trabajo con la que subsanar el déficit al que se hacía referencia anteriormente

Potencialidad de la modelización state-space y el filtro de Kalman para el anàlisis regional : una aplicación para el índice de actividad industrial

El déficit existente a nuestro país con respecto a la disponibilidad de indicadores cuantitativos con los que llevar a término un análisis coyuntural de la actividad industrial regional ha abierto un debate centrado en el estudio de cuál es la metodología más adecuada para elaborar indicadores de estas características. Dentro de este marco, en este trabajo se presentan las principales conclusiones obtenidas en anteriores estudios (Clar, et. al., 1997a, 1997b y 1998) sobre la idoneidad de extender las metodologías que actualmente se están aplicando a las regiones españolas para elaborar indicadores de la actividad industrial mediante métodos indirectos. Estas conclusiones llevan a plantear una estrategia distinta a las que actualmente se vienen aplicando. En concreto, se propone (siguiendo a Israilevich y Kuttner, 1993) un modelo de variables latentes para estimar el indicador de la producción industrial regional. Este tipo de modelo puede especificarse en términos de un modelo statespace y estimarse mediante el filtro de Kalman. Para validar la metodología propuesta se estiman unos indicadores de acuerdo con ella para tres de las cuatro regiones españolas que disponen d¿un Índice de Producción Industrial (IPI) elaborado mediante el método directo (Andalucía, Asturias y el País Vasco) y se comparan con los IPIs publicados (oficiales). Los resultados obtenidos muestran el buen comportamiento de l¿estrategia propuesta, abriendo así una línea de trabajo con la que subsanar el déficit al que se hacía referencia anteriormente

Adaptive Markov chain Monte Carlo and Bayesian filtering for state space models

This thesis is concerned with the state and parameter estimation in state space models. The estimation of states and parameters is an important task when mathematical modeling is applied to many different application areas such as the global positioning systems, target tracking, navigation, brain imaging, spread of infectious diseases, biological processes, telecommunications, audio signal processing, stochastic optimal control, machine learning, and physical systems. In Bayesian settings, the estimation of states or parameters amounts to computation of the posterior probability density function. Except for a very restricted number of models, it is impossible to compute this density function in a closed form. Hence, we need approximation methods. A state estimation problem involves estimating the states (latent variables) that are not directly observed in the output of the system. In this thesis, we use the Kalman filter, extended Kalman filter, Gauss–Hermite filters, and particle filters to estimate the states based on available measurements. Among these filters, particle filters are numerical methods for approximating the filtering distributions of non-linear non-Gaussian state space models via Monte Carlo. The performance of a particle filter heavily depends on the chosen importance distribution. For instance, inappropriate choice of the importance distribution can lead to the failure of convergence of the particle filter algorithm. In this thesis, we analyze the theoretical Lᵖ particle filter convergence with general importance distributions, where p ≥2 is an integer. A parameter estimation problem is considered with inferring the model parameters from measurements. For high-dimensional complex models, estimation of parameters can be done by Markov chain Monte Carlo (MCMC) methods. In its operation, the MCMC method requires the unnormalized posterior distribution of the parameters and a proposal distribution. In this thesis, we show how the posterior density function of the parameters of a state space model can be computed by filtering based methods, where the states are integrated out. This type of computation is then applied to estimate parameters of stochastic differential equations. Furthermore, we compute the partial derivatives of the log-posterior density function and use the hybrid Monte Carlo and scaled conjugate gradient methods to infer the parameters of stochastic differential equations. The computational efficiency of MCMC methods is highly depend on the chosen proposal distribution. A commonly used proposal distribution is Gaussian. In this kind of proposal, the covariance matrix must be well tuned. To tune it, adaptive MCMC methods can be used. In this thesis, we propose a new way of updating the covariance matrix using the variational Bayesian adaptive Kalman filter algorithm.

La structure de Jordan des matrices de transfert des modèles de boucles et la relation avec les hamiltoniens XXZ

Les modèles sur réseau comme ceux de la percolation, d’Ising et de Potts servent à décrire les transitions de phase en deux dimensions. La recherche de leur solution analytique passe par le calcul de la fonction de partition et la diagonalisation de matrices de transfert. Au point critique, ces modèles statistiques bidimensionnels sont invariants sous les transformations conformes et la construction de théories des champs conformes rationnelles, limites continues des modèles statistiques, permet un calcul de la fonction de partition au point critique. Plusieurs chercheurs pensent cependant que le paradigme des théories des champs conformes rationnelles peut être élargi pour inclure les modèles statistiques avec des matrices de transfert non diagonalisables. Ces modèles seraient alors décrits, dans la limite d’échelle, par des théories des champs logarithmiques et les représentations de l’algèbre de Virasoro intervenant dans la description des observables physiques seraient indécomposables. La matrice de transfert de boucles D_N(λ, u), un élément de l’algèbre de Temperley- Lieb, se manifeste dans les théories physiques à l’aide des représentations de connectivités ρ (link modules). L’espace vectoriel sur lequel agit cette représentation se décompose en secteurs étiquetés par un paramètre physique, le nombre d de défauts. L’action de cette représentation ne peut que diminuer ce nombre ou le laisser constant. La thèse est consacrée à l’identification de la structure de Jordan de D_N(λ, u) dans ces représentations. Le paramètre β = 2 cos λ = −(q + 1/q) fixe la théorie : β = 1 pour la percolation et √2 pour le modèle d’Ising, par exemple. Sur la géométrie du ruban, nous montrons que D_N(λ, u) possède les mêmes blocs de Jordan que F_N, son plus haut coefficient de Fourier. Nous étudions la non diagonalisabilité de F_N à l’aide des divergences de certaines composantes de ses vecteurs propres, qui apparaissent aux valeurs critiques de λ. Nous prouvons dans ρ(D_N(λ, u)) l’existence de cellules de Jordan intersectorielles, de rang 2 et couplant des secteurs d, d′ lorsque certaines contraintes sur λ, d, d′ et N sont satisfaites. Pour le modèle de polymères denses critique (β = 0) sur le ruban, les valeurs propres de ρ(D_N(λ, u)) étaient connues, mais les dégénérescences conjecturées. En construisant un isomorphisme entre les modules de connectivités et un sous-espace des modules de spins du modèle XXZ en q = i, nous prouvons cette conjecture. Nous montrons aussi que la restriction de l’hamiltonien de boucles à un secteur donné est diagonalisable et trouvons la forme de Jordan exacte de l’hamiltonien XX, non triviale pour N pair seulement. Enfin nous étudions la structure de Jordan de la matrice de transfert T_N(λ, ν) pour des conditions aux frontières périodiques. La matrice T_N(λ, ν) a des blocs de Jordan intrasectoriels et intersectoriels lorsque λ = πa/b, et a, b ∈ Z×. L’approche par F_N admet une généralisation qui permet de diagnostiquer des cellules intersectorielles dont le rang excède 2 dans certains cas et peut croître indéfiniment avec N. Pour les blocs de Jordan intrasectoriels, nous montrons que les représentations de connectivités sur le cylindre et celles du modèle XXZ sont isomorphes sauf pour certaines valeurs précises de q et du paramètre de torsion v. En utilisant le comportement de la transformation i_N^d dans un voisinage des valeurs critiques (q_c, v_c), nous construisons explicitement des vecteurs généralisés de Jordan de rang 2 et discutons l’existence de blocs de Jordan intrasectoriels de plus haut rang.

Sparse model identification using orthogonal forward regression with basis pursuit and D-optimality

An efficient model identification algorithm for a large class of linear-in-the-parameters models is introduced that simultaneously optimises the model approximation ability, sparsity and robustness. The derived model parameters in each forward regression step are initially estimated via the orthogonal least squares (OLS), followed by being tuned with a new gradient-descent learning algorithm based on the basis pursuit that minimises the l(1) norm of the parameter estimate vector. The model subset selection cost function includes a D-optimality design criterion that maximises the determinant of the design matrix of the subset to ensure model robustness and to enable the model selection procedure to automatically terminate at a sparse model. The proposed approach is based on the forward OLS algorithm using the modified Gram-Schmidt procedure. Both the parameter tuning procedure, based on basis pursuit, and the model selection criterion, based on the D-optimality that is effective in ensuring model robustness, are integrated with the forward regression. As a consequence the inherent computational efficiency associated with the conventional forward OLS approach is maintained in the proposed algorithm. Examples demonstrate the effectiveness of the new approach.

Neurofuzzy control using Kalman filtering state feedback with coloured noise for unknown non-linear processes

This paper presents a controller design scheme for a priori unknown non-linear dynamical processes that are identified via an operating point neurofuzzy system from process data. Based on a neurofuzzy design and model construction algorithm (NeuDec) for a non-linear dynamical process, a neurofuzzy state-space model of controllable form is initially constructed. The control scheme based on closed-loop pole assignment is then utilized to ensure the time invariance and linearization of the state equations so that the system stability can be guaranteed under some mild assumptions, even in the presence of modelling error. The proposed approach requires a known state vector for the application of pole assignment state feedback. For this purpose, a generalized Kalman filtering algorithm with coloured noise is developed on the basis of the neurofuzzy state-space model to obtain an optimal state vector estimation. The derived controller is applied in typical output tracking problems by minimizing the tracking error. Simulation examples are included to demonstrate the operation and effectiveness of the new approach.

Estrutura a termo da taxa de juros no Brasil e previsibilidade de ciclos econômicos

O objetivo deste trabalho é caracterizar a Curva de Juros Mensal para o Brasil através de três fatores, comparando dois tipos de métodos de estimação: Através da Representação em Espaço de Estado é possível estimá-lo por dois Métodos: Filtro de Kalman e Mínimos Quadrados em Dois Passos. Os fatores têm sua dinâmica representada por um Modelo Autorregressivo Vetorial, VAR(1), e para o segundo método de estimação, atribui-se uma estrutura para a Variância Condicional. Para a comparação dos métodos empregados, propõe-se uma forma alternativa de compará-los: através de Processos de Markov que possam modelar conjuntamente o Fator de Inclinação da Curva de Juros, obtido pelos métodos empregados neste trabalho, e uma váriavel proxy para Desempenho Econômico, fornecendo alguma medida de previsão para os Ciclos Econômicos.

Proposal of an alternative transmission line model for symmetrical and asymmetrical configurations

This article shows a transmission line model for simulation of fast and slow transients, applied to symmetrical or asymmetrical configurations. A transmission line model is developed based on lumped elements representation and state-space techniques. The proposed methodology represents a practical procedure to model three-phase transmission lines directly in time domain, without the explicit or implicit use of inverse transforms. In three-phase representation, analysis modal techniques are applied to decouple the phases in their respective propagation modes, using a correction procedure to set a real and constant matrix for untransposed lines with or without vertical symmetry plane. The proposed methodology takes into account the frequency-dependent parameters of the line and in order to include this effect in the state matrices, a fitting procedure is applied. To verify the accuracy of the proposed state-space model in frequency domain, a simple methodology is described based on line distributed parameters and transfer function associated with input/output signals of the lumped parameters representation. In addition, this article proposes the use of a fast and robust integration procedure to solve the state equations, enabling transient and steady-state simulations. The results obtained by the proposed methodology are compared with several established transmission line models in EMTP, taking into account an asymmetrical three-phase transmission line. The principal contribution of the proposed methodology is to handle a steady fundamental signal mixed with fast and slow transients, including impulsive and oscillatory behavior, by a practical procedure applied directly in time domain for symmetrical or asymmetrical representations. (C) 2011 Elsevier Ltd. All rights reserved.

Stabilizability and Disturbance Rejection with State-Derivative Feedback

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Controle robusto com realimentação derivativa de sistemas não lineares via LMI

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)