SAFE: Spectral Evolution Analysis Feature Extraction for Non-Stationary Time Series Prediction
This addresses the problem of adapting to non-stationary data in real-time prediction for applications like forecasting, though it appears incremental as it builds on existing machine learning methods.
The paper tackles non-stationarity in time series prediction by proposing SAFE, a method that detects changes in spectral content to trigger adaptation in online predictors, and it shows significant computational savings while maintaining high prediction performance.
This paper presents a practical approach for detecting non-stationarity in time series prediction. This method is called SAFE and works by monitoring the evolution of the spectral contents of time series through a distance function. This method is designed to work in combination with state-of-the-art machine learning methods in real time by informing the online predictors to perform necessary adaptation when a non-stationarity presents. We also propose an algorithm to proportionally include some past data in the adaption process to overcome the Catastrophic Forgetting problem. To validate our hypothesis and test the effectiveness of our approach, we present comprehensive experiments in different elements of the approach involving artificial and real-world datasets. The experiments show that the proposed method is able to significantly save computational resources in term of processor or GPU cycles while maintaining high prediction performances.