Prediction Model Selection with Frequency Check on Decomposed Time Series

High prediction accuracies at time series modeling and forecasting is of the utmost importance for a variety of application domains. Various time series prediction methods exist that use linear and nonlinear models separately or combination of both. These methods highly increase prediction performance results when they are applied on a many number of stationary components obtained by more sophisticated decomposition techniques. Although these stationary components are easily predictable, they each have different characteristics. In this study, we have developed a hybrid method that aims to increase prediction performance in time series considering these differences. The developed method decomposes given time series into many stationary components with two-level decomposition using Movingaverage (MA) filter and the Empirical Mode Decomposition (EMD) techniques. Then, the obtained components are modeled separately by appropriate methods according to the frequency changing rates calculated by fourier analysis. The evaluation of the developed method is performed on three different publicly available benchmark dataset and achieved highly successful results as compared to existing well-accepted hybrid methods.