Time series prediction with a weighted bidirectional multi-stream extended Kalman filter

Authors:
Xiao Hu;Danil V. Prokhorov;Donald C. Wunsch II
Affiliations:
One Research Circle, General Electrical Global Research Center, Niskayuna, NY 12309, USA;Toyota Technical Center, Ann Arbor, MI 48105, USA;Applied Computational Intelligence Lab, Department of Electrical & Computer Engineering, University of Missouri-Rolla, Rolla, MO 65409, USA
Venue:
Neurocomputing
Year:
2007

Citing 5
Cited 2

The roots of backpropagation: from ordered derivatives to neural networks and political forecasting

The roots of backpropagation: from ordered derivatives to neural networks and political forecasting
Gradient-based learning algorithms for recurrent networks and their computational complexity

Backpropagation
Adaptive filter theory (3rd ed.)

Adaptive filter theory (3rd ed.)
Comparative study of stock trend prediction using time delay, recurrent and probabilistic neural networks

IEEE Transactions on Neural Networks
Neurocontrol of nonlinear dynamical systems with Kalman filter trained recurrent networks

IEEE Transactions on Neural Networks

A novel approach for distributed application scheduling based on prediction of communication events

Future Generation Computer Systems
On the dynamic evidential reasoning algorithm for fault prediction

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.01

Visualization

Abstract

We use a multi-stream extended Kalman filter for the CATS benchmark (Competition on Artificial Time Series), to train recurrent multilayer perceptrons. A weighted bidirectional approach is adopted to combine forward and backward predictions and to generate the final predictions on the missing points.