
The authors propose a deep learning-based approach, called DeepVARwT, for modeling and forecasting multivariate time series with trends. The method simultaneously estimates the trend and the dependence structure using a Long Short-Term Memory (LSTM) network and a vector autoregressive (VAR) model.


coremsg

deep-learning-for-modeling-and-forecasting-multivariate-time-series-with-trends


Deep Learning for Modeling and Forecasting Multivariate Time Series with Trends
