This paper describes the use of a multi-stream extended Kalman filter (EKF) to tackle the IJCNN 2004 challenge problem - time series prediction on CATS benchmark. A weighted bidirectional approach was adopted in the experiments to incorporate the forward and backward predictions of the time series. EKF is a practical, general approach to neural networks training. It consists of the following: 1) gradient calculation by backpropagation through time (BPTT); 2) weight updates based on the extended Kalman filter; and 3) data presentation using multi-stream mechanics.

Meeting Name

IEEE International Joint Conference on Neural Networks, 2004


Electrical and Computer Engineering

Keywords and Phrases

CATS Benchmark; EKF; IJCNN 2004 Challenge Problem; Kalman Filters; Backpropagation; Backpropagation Through Time; Competition on Artificial Time Series; Data Presentation; Filtering Theory; Gradient Calculation; Gradient Methods; Multistream Extended Kalman Filter; Multistream Mechanics; Neural Nets; Neural Networks Training; Prediction Theory; Time Series; Time Series Prediction; Weighted Bidirectional Method

International Standard Serial Number (ISSN)


Document Type

Article - Conference proceedings

Document Version

Final Version

File Type





© 2004 Institute of Electrical and Electronics Engineers (IEEE), All rights reserved.

Publication Date

01 Jan 2004