论文信息 - Multi-stage Algorithm Based on Neural Network Committee for Prediction and Search for Precursors in Multi-dimensional Time Series

Multi-stage Algorithm Based on Neural Network Committee for Prediction and Search for Precursors in Multi-dimensional Time Series

The studied problem is prediction of time series based on preceding values of several time series (a multi-dimensional time series). Besides prediction itself, the task is finding precursors, i.e. determination of a set of the most significant input features in coordinates "initial time series --- lag". A four-stage prediction algorithm based on neural network committee has been suggested, implemented and studied. The algorithm has been successfully tested on one model problem and on one real world problem.

[1] David H. Wolpert,et al. Stacked generalization , 1992, Neural Networks.

[2] H. Gleisner,et al. Predicting geomagnetic storms from solar-wind data using time-delay neural networks , 1996 .

[3] Yu. V. Orlov,et al. Neural Network Algorithm for Events Forecasting and Its Application to Space Physics Data , 2005, ICANN.

[4] Erkki Oja,et al. Artificial Neural Networks: Formal Models and Their Applications - ICANN 2005, 15th International Conference, Warsaw, Poland, September 11-15, 2005, Proceedings, Part II , 2005, International Conference on Artificial Neural Networks.

[5] Yu. V. Orlov,et al. Neural network algorithms for analyzing multidimensional time series for predicting events and their application to study of Sun-Earth relations , 2007, Pattern Recognition and Image Analysis.