English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Zendy Plus

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Zendy Tools

Zendy Open

Multivariate time series forecasting is of great importance to manyscientific disciplines and industrial sectors. The evolution of a multivariatetime series depends on the dynamics of its variables and the connectivitynetwork of causal interrelationships among them. Most of the existing timeseries models do not account for the causal effects among the system'svariables and even if they do they rely just on determining thebetween-variables causality network. Knowing the structure of such a complexnetwork and even more specifically knowing the exact lagged variables thatcontribute to the underlying process is crucial for the task of multivariatetime series forecasting. The latter is a rather unexplored source ofinformation to leverage. In this direction, here a novel neural network-basedarchitecture is proposed, termed LAgged VAriable Representation NETwork(LAVARNET), which intrinsically estimates the importance of lagged variablesand combines high dimensional latent representations of them to predict futurevalues of time series. Our model is compared with other baseline and state ofthe art neural network architectures on one simulated data set and four realdata sets from meteorology, music, solar activity, and finance areas. Theproposed architecture outperforms the competitive architectures in most of theexperiments.

LAVARNET: Neural Network Modeling of Causal Variable Relationships for Multivariate Time Series Forecasting