Abstract | ||
---|---|---|
Bounding ellipsoid (BE) algorithms offer an attractive alternative to traditional training algorithms for neural networks, for example, backpropagation and least squares methods. The benefits include high computational efficiency and fast convergence speed. In this paper, we propose an ellipsoid propagation algorithm to train the weights of recurrent neural networks for nonlinear systems identification. Both hidden layers and output layers can be updated. The stability of the BE algorithm is proven. |
Year | DOI | Venue |
---|---|---|
2009 | 10.1109/TNN.2009.2015079 | IEEE Transactions on Neural Networks |
Keywords | Field | DocType |
lyapunov-like technique,neural network,ellipsoid algorithm,identification,hidden layer,learning (artificial intelligence),high computational efficiency,nonlinear systems identification,recurrent neural networks,convergence speed,nonlinear systems,ellipsoid propagation algorithm,stable bounding ellipsoid algorithm,traditional training algorithm,bounding ellipsoid (be),bounding ellipsoid,recurrent neural network training,attractive alternative,recurrent neural nets,recurrent neural network,recurrent neural networks training,lyapunov methods,neural networks,convergence,learning artificial intelligence,algorithms,feedback,backpropagation,ellipsoids,function approximation,indexing terms,least square method,feedforward neural networks,computer simulation,uncertainty | Feedforward neural network,Ellipsoid,Pattern recognition,Computer science,Recurrent neural network,Artificial intelligence,System identification,Backpropagation,Artificial neural network,Ellipsoid method,Machine learning,Bounding overwatch | Journal |
Volume | Issue | ISSN |
20 | 6 | 1941-0093 |
Citations | PageRank | References |
16 | 0.81 | 18 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Wen Yu | 1 | 283 | 22.70 |
José De Jesús Rubio | 2 | 574 | 36.29 |