Various Type of Wavelet Filters on Time Series Forecasting
Keun-Tae Park, Jun‐Geol Baek · 2017
Forecasting time series data is one of the most important subjects that is useful and applicable in real life. The objective of this study improves the performances of time series forecasting method called ARIMA with wavelet transform. The proposed method is taking an optimal type of Daubechies wavelet transform functions. Real case datasets in existing paper are used to compare the performance with original and existing forecasting methods. The results of experiment demonstrate the usefulness and superiority of the proposed method with more possibility.