(Institute of Industrial Science, The University of Tokyo) Scientists at the University of Tokyo and Kozo Keikaku Engineering Inc. have shown how to combine the forecasts of a collection of suboptimal ‘delay embedding’ predictors for time series data. This work may help improve the forecasting of floods, stock market gyrations, spatio-temporal brain dynamics, and ecological resource fluctuations

Original source: https://www.eurekalert.org/pub_releases/2020-01/iois-wot011620.php