1,721,044 research outputs found
I minimi quadrati a due stadi per la stima di serie spazio-temporali con effetto simultaneo
In a general formulation of a spatio-temporal models we believe that simultaneous spatial effects cannot be disregarded. If simultaneous spatio-temporal effects are present in the data generating process, LS estimates become inconsistent. To estimate this effects, recent literature propose the use of maximum likelihood estimates, while in our opinion two-stage least square seems to be more appropriate and computationally easier to handle.
In this paper we estimate a model in which the specification of space-time simultaneous effects take into account intensity and direction of relations among variables in different sites not depending on distance. An empirical application to monthly data on the index of industrial turnover in Lombardy is presented
A dynamic model for solar radiation in the Tyrrhenian Basin of Central Mediterranean area
Studying the level of solar radiation is important for problems related to both environmental pollution and alternative energy development. In this work a space-time model for solar radiation in the Tyrrhenian basin is presented. Three main features of the model must be stressed because of their importance in modelling space-time variability of a phenomenon. The first and most important one is that relations between solar radiation in different sites are an outcome of the model’s estimation procedure. With this approach spatial weights are not bound to be symmetrical and proportional to distance between locations or to be constant over time. The second one is the presence of a simultaneous effect among locations as the solar radiation in one of them is a function of what simultaneously happens in all the other ones. The third main feature of the model is represented by constrained estimation on the basis of a priori knowledge about the phenomenon that allows to cope the problem of the increased number of parameters
Spatio-Temporal weight with simultaneous effect for environmental data
In this work we present space-time models with simultaneous effect in which the relations between
phenomenon measured in different sites are not determined on the base of an a priori structure, but
are estimated in the model’s estimation procedure.
This approach avoids superimposition of a priori space fixed structure so that the spatial weights are
not bound at being symmetrical and have not to be constant over time. In this way, spatial weights will
then take into account the intensity and the direction of influence that a single station, in a given time
including the contemporaneous one, has on the whole spatial system and vice versa.
To reduce the increase in the number of parameters a procedure for constrained estimation is
presented
Full Information Least Orthogonal Distance Estimator of structural parameters in simultaneous equation models
An extension of Limited Information Least Orthogonal Distance Estimator (LI LODE) for structural parameters of simultaneous equation models, to a Full Information context, is presented. The proposed extension is based on characteristic roots and vectors of a matrix deriving from the so called over-identifying restrictions
A multiple bi-dimensional BEKK model for portfolio volatility Matrix estimation of the Italian stock market
- …
