A Novel Approach for Seasonality and Trend Detection using Fast Fourier Transform in Box-Jenkins Algorithm
Authors:
- 30 August 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Forecasting is the first step to deal with the new generation of renewable energy systems. The accuracy of the forecasting techniques is very important. Time series technique is one of the powerful tools used for forecasting, but it works well with stationary data. In addition, non-stationary time series can cause unexpected behaviors or create a non-existing relationship between two variables. This work was motivated by the need of detecting the seasonality and trend for a given data. The trend and seasonality components are very important in dealing with forecasting. Based on the trend and seasonality we could use clustered regions and feed the clustered regions to a forecasting technique like ANN, WNN or Kalman Filtering. Then aggregating the forecasted data back again for a better performance. In this paper we use Fast Fourier Transform (FFT) in Box-Jenkins approach instead of Autocorrelation Function (ACF) for the seasonality and trend detection. The present study is validated using visual inspection, statistical tests and time series decomposition in identifying the trend and the seasonality by applying them to wind speed time series. The results of FFT technique and ACF are compared to the results of the most well-known techniques. The results show that some methods have minor limitations in determining either the trend or the seasonality compared to FFT.Keywords
This publication has 15 references indexed in Scilit:
- A proposed intelligent short-term load forecasting hybrid models of ANN, WNN and KF based on clustering techniques for smart gridElectric Power Systems Research, 2020
- A novel approach for harmonic tidal currents constitutions forecasting using hybrid intelligent models based on clustering methodologiesRenewable Energy, 2020
- Identifying Seasonality in Time Series by Applying Fast Fourier Transform2019
- Local short term electricity load forecasting: Automatic approaches2017
- ARIMA modeling of wind speed for wind farm reliability analysis2014
- A Proposed ANN and FLSM Hybrid Model for Tidal Current Magnitude and Direction ForecastingIEEE Journal of Oceanic Engineering, 2013
- Very Short-Term Load Forecasting Based on ARIMA Model and Intelligent Systems2009
- ARIMA models to predict next-day electricity pricesIEEE Transactions on Power Systems, 2003
- Stationarity and nonstationarity in time series analysisPhysica D: Nonlinear Phenomena, 1996
- Visual inspection of data: does the eyeball fit the trend?1994