Turning mechanistic models into forecasters by using machine learning
This work addresses the challenge of forecasting in systems with non-stationary dynamics, such as epidemiology or environmental monitoring, though it is incremental by extending existing data-driven discovery methods.
The authors tackled the problem of modeling complex dynamical systems with evolving dynamics by introducing time-varying parameters into data-driven equation discovery, achieving a mean absolute error below 3% for learning and below 6% for forecasting up to a month ahead. They validated their model on datasets like SIR and greenhouse gas concentration, outperforming CNN-LSTM and GBM methods.
The equations of complex dynamical systems may not be identified by expert knowledge, especially if the underlying mechanisms are unknown. Data-driven discovery methods address this challenge by inferring governing equations from time-series data using a library of functions constructed from the measured variables. However, these methods typically assume time-invariant coefficients, which limits their ability to capture evolving system dynamics. To overcome this limitation, we allow some of the parameters to vary over time, learn their temporal evolution directly from data, and infer a system of equations that incorporates both constant and time-varying parameters. We then transform this framework into a forecasting model by predicting the time-varying parameters and substituting these predictions into the learned equations. The model is validated using datasets for Susceptible-Infected-Recovered, Consumer--Resource, greenhouse gas concentration, and Cyanobacteria cell count. By dynamically adapting to temporal shifts, our proposed model achieved a mean absolute error below 3\% for learning a time series and below 6\% for forecasting up to a month ahead. We additionally compare forecasting performance against CNN-LSTM and Gradient Boosting Machine (GBM), and show that our model outperforms these methods across most datasets. Our findings demonstrate that integrating time-varying parameters into data-driven discovery of differential equations improves both modeling accuracy and forecasting performance.