Self-supervised learning for fast and scalable time series hyper-parameter tuning
This addresses the need for fast and scalable hyper-parameter tuning in time series analysis, which is incremental as it builds on existing methods by replacing search with learning.
The paper tackles the problem of computationally expensive hyper-parameter tuning for time series models by proposing a self-supervised learning framework (SSL-HPT) that uses time series features to directly output optimal hyper-parameters, achieving speeds 6-20 times faster than search-based methods while maintaining comparable forecasting accuracy.
Hyper-parameters of time series models play an important role in time series analysis. Slight differences in hyper-parameters might lead to very different forecast results for a given model, and therefore, selecting good hyper-parameter values is indispensable. Most of the existing generic hyper-parameter tuning methods, such as Grid Search, Random Search, Bayesian Optimal Search, are based on one key component - search, and thus they are computationally expensive and cannot be applied to fast and scalable time-series hyper-parameter tuning (HPT). We propose a self-supervised learning framework for HPT (SSL-HPT), which uses time series features as inputs and produces optimal hyper-parameters. SSL-HPT algorithm is 6-20x faster at getting hyper-parameters compared to other search based algorithms while producing comparable accurate forecasting results in various applications.