Advanced Forecasting Methods Based on Spectral Analysis

This article presents a time series estimation and prediction methods with the use of classic and advanced forecasting tools. In addition, rarely applied in practice approach using spectral analysis to an identification of variation patterns and prediction will be presented. The effect of spectral analysis will be an estimation of prediction model parameters. The main assumption is the model consist of trigonometric functions combination with a certain frequency. The model includes only those frequencies which have greatest influence on process variation. The effectiveness of the method will be examined by a numerical example. The area of the proposed methodology is broad and goes beyond economics.