Designing an Associative Memory via Optimal Training for Fault Diagnosis

Author

Ruz-Hernandez, Jose A. ; Sanchez, Edgar N. ; Suarez, Dionisio A.

Author_Institution

Univ. Autonoma del Carmen, Campeche

fYear

0

fDate

0-0 0

Firstpage

4338

Lastpage

4345

Abstract

In this paper, the authors discuss a new synthesis approach to train associative memories, based on recurrent neural networks. They propose to determine the weight vector as the optimal solution of a linear combination of support patterns. The proposed training algorithm maximizes the margin between the training patterns and the decision boundary. The design problem considers: (1) obtaining of weights via an optimal hyperplane algorithm for SVMs and (2) obtaining conditions to reduce the total number of spurious memories. This algorithm is applied to the synthesis of an associative memory for fault diagnosis in fossil electric power plants.

Keywords

content-addressable storage; fault diagnosis; power generation faults; power system analysis computing; recurrent neural nets; support vector machines; associative memory; fault diagnosis; optimal hyperplane algorithm; optimal training; recurrent neural network; support vector machine; Algorithm design and analysis; Associative memory; Design methodology; Fault diagnosis; Machine learning algorithms; Network synthesis; Neural networks; Quadratic programming; Recurrent neural networks; Support vector machines; Associative memory; fault diagnosis; fossil electric power plants; optimal hyperplane; recurrent neural networks; support vector machines;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks, 2006. IJCNN '06. International Joint Conference on

Conference_Location

Vancouver, BC

Print_ISBN

0-7803-9490-9

Type

conf

DOI

10.1109/IJCNN.2006.247010

Filename

1716699