Improved Facebook Prophet Model Using Singular Spectrum Analysis for Short-Term Load Forecasting
- Resource Type
- Conference
- Authors
- Saini, Priyesh; Abhinav, Kumar; Parida, S. K.
- Source
- 2023 IEEE International Conference on Energy Technologies for Future Grids (ETFG) Energy Technologies for Future Grids (ETFG), 2023 IEEE International Conference on. :1-6 Dec, 2023
- Subject
- Power, Energy and Industry Applications
Analytical models
Social networking (online)
Predictive models
Market research
Feature extraction
Spectral analysis
Load modeling
Facebook prophet
Singular Spectrum Analysis (SSA)
stationarity
- Language
This study proposes a unique approach that combines feature selection and Singular Spectrum Analysis (SSA), to improve the precision of Facebook Prophet model for short-term load forecasting (STLF). Correlation analysis is carried out to determine the key input features related to load demand. The non-stationary behavior of time series is addressed by applying SSA to extract the trend component that captures the long-term trends and changes in load demand. To determine the efficacy of the suggested approach on test data, a variety of evaluation metrics are used, and a thorough comparison with the existing prophet model is made to highlight the significance of SSA in improving forecasting accuracy.