Application of methods of graph theory in the problem of estimation of parameters of a regression model in the presence of constraints of a special type of constraints

The paper is concerned with a regression model used in isolation of anomalous events in seismic data. The model parameter identification is interpreted in terms of graph theory. In this way the kernel structure of the regression matrix is studied. To improve the consistence and reduce the dimension of the equation set, constraints of a special kind are imposed. A graph-theoretic procedure is developed for computing the LMS estimate with these constraints.