Econometric Forecasting Models Based on Forecast Combination Methods

In this work econometric forecasting models which increase the accuracy of the forecast are introduced on the basis of the methodology of combined forecast and wavelet analysis. Time series forecasting models encompass many assumptions for their successful implementation. For instance, regression analysis methods assume what curve can describe a time series; in autoregressive models, it is assumed that the value of the current indicator linearly depends on some quantity of its lag values. However, these assumptions are rarely met and it weakens the predictive capability of a model. More accurate models are needed in order to avoid that.