Linear modeling and the coherence function

摘要: In this paper, a tutorial introduction to the magnitude squared (MS) coherence function, and its relation linear modeling, is interleaved with new methods for characterizing bidirectional unidirectional association between two time series. A method based on concept of an MS sequence which inverse z-transform function. The function approximated as single real-valued ratio polynominals in z. determined by matrix equation whose entries are convolutions correlation functions. Solutions corresponding parameters employ eigenvalue-eigenvector decomposition. algorithm compared traditional periodogram-based using computer simulations. simulations also demonstrate effects parameter overdetermination overordering. sensitivity equations computed via numerical algebra condition number.