Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
418113 | Computational Statistics & Data Analysis | 2007 | 11 Pages |
A new approach to the description of national economies is presented. For this the state space viewpoint is used, which is utilized mostly in the theory of dynamical systems and in control theory. Gross domestic product, inflation, and unemployment rates are taken as state variables. It is demonstrated that for the period of time under consideration the phase trajectory of each of the V4 countries (Slovak Republic, Czech Republic, Hungary, and Poland) lies approximately in one plane, so that the economic development of each country can be associated with a corresponding plane in the state space. The suggested approach opens a way to a new set of economic indicators (for example, normal vectors of national economies, various plane slopes, 2D angles between the planes corresponding to different economies, etc.). The tool used for computations is orthogonal regression (alias orthogonal distance regression, alias total least squares method), and general arguments for using orthogonal regression instead of the classical regression based on the least squares method are given. A MATLAB routine for fitting 3D data to lines and planes in 3D is provided.