Application of Prony-transform for time series modeling in economics

In article the decision of problems of modelling and the forecast of numerical numbers with application of spectral estimation methodology is considered. Use of qualitative models of data allows to carry out the authentic forecast that is an actual problem at acceptance of administrative decisions. The forecast methodology based on a Prony-transform, allowing to represent an initial numerical number in the form of new of some - to some linear combination exponential functions. Generally complex member of the series are defined by amplitude, frequency, a phase and attenuation. In work the basic concepts and parities of Prony-transform are resulted, the behaviour separate a component of transformation for typical cases is analyzed. Features of calculation of Prony-spectrum and its differences from Fourier-spectrum are considered. Separate stages of work of algorithm, a problem of specific realization are discussed. With application of the described procedure and on the basis of the program written by the author the model of monthly variation of a consumer price index in Russia for some last years is under construction, the possibility of the forecast on the basis of such model is shown.