Journal:Informatica
Volume 3, Issue 1 (1992), pp. 37–46
Abstract
The dynamic programming method for estimation of many change-points in univariate autoregressive (AR) sequences with known AR parameters between change-points is investigated. A problem how to use this method for long autoregressive sequences is solved and a constructive solution is given. A simulation experiment illustrates the advantages of the solution obtained.
Journal:Informatica
Volume 3, Issue 1 (1992), pp. 21–36
Abstract
The idea of predicting the case, when the considered long-term ARMA model, fitted to the observed time series tends to become unstable because of deep changes in the structural stability of data, is developed in this paper. The aim is to predict a possible unstable regime of the process {Xt,t∈T}τ-steps in advance before it will express itself by a high level crossing or large variance of an output variable Xt. The problem is solved here for locally stationary AR(p) sequences {Xt,t∈T}, whose estimated parameters can reach critical sets located at the boundary of the stability area. An alarm function and an alarm set are fitted here to predict catastrophic failures in systems output τ units in advance for given τ>0 and a confidence level γ. The probability of false alarm is derived explicitly for AR(1) depending on τ,γ and N – the number of the last observations of {Xt}.
Journal:Informatica
Volume 3, Issue 1 (1992), pp. 3–20
Abstract
We present a new method for solving the change-point detection problem for ARMA systems which are assumed to have a slow and non-decaying drift after the change occurs. The proposed technique is inspired by the stochastic complexity theory, which gives a basis of comparison of different models with different change-point times. Some partial results on the analysis of the estimator are stated. A simulation is included which shows that the approach exhibits surprisingly good detection capabilities.
Journal:Informatica
Volume 2, Issue 4 (1991), pp. 579–592
Abstract
In the previous paper (Pupeikis, 1990) the problem of model order determination in the presence of outliers in observations has been considered. The aim of the given paper is the development of the recursive algorithms of computation of M-estimates ensuring their stability conditions. In this connection the approach, based on adaptive Huber's monotone psi-function, is worked out. It is also used for the detection of the outliers in time series and for the correction both outliers and M-estimates during successive calculations. The results of numerical simulation by computer (Fig. 1 and Table 1) are given.
Journal:Informatica
Volume 2, Issue 4 (1991), pp. 564–578
Abstract
The present paper considers a constant-parameter estimation algorithm for an analog transfer function by discrete observations of the object's input and output variables. The algorithm is based on supplementary variables and least-squares methods. It is assumed, that the order of the transfer function is known and the derivatives of the input and output variables are non-measurable. The supplementary variables and their derivatives are constructed from discrete observations of the input and output variables by applying a numerically realized analog filter. Investigation results for the estimate properties are presented. The results were obtained by the method of statistical simulation.
Journal:Informatica
Volume 2, Issue 4 (1991), pp. 552–563
Abstract
The aim of this paper is to concentrate in one place and to show the relations among the matrix block impulse response, block impulse response, matrix impulse response and impulse response of linear time-varying (LTV) systems, frozen-time LTV systems, linear periodically time-varying (LPTV) systems, and linear time-invariant (LTI) systems.
Journal:Informatica
Volume 2, Issue 4 (1991), pp. 539–551
Abstract
The local optimization techniques is the basis of majority of regular (exact) algorithms for the non-monoton pseudoboolean functions optimization as the most simple and, accordingly, the most universal method of the discrete optimization. However, the local optimization method does not guarantee the elimination of the total examination when the pseudoboolean optimization problem in a general state is solved. In the present paper the cutting off algorithms are suggested which guarantee the total examination elimination for any pseudoboolean optimization problem.