Abstract
In Telephone network traffic management, real-time disturbance detection and identification is a complex problem because of the large variability in both temporal and spatial characteristics even from the same disturbance. This paper presents the design and evaluation of a diagnosis system based on a combination of the discriminative powers of multiple classifiers. Several different Neural Networks were initially trained to perform the same task. Combinations of these different pattern classifiers are examined and comparedto the performance of a single network. The underlying idea is that the combination of multiple classifiers promises to improve the performance and fault tolerance. Combinations are considered as linear and non-linear weighted rules. We demonstrate our stztdy on a model long distance telephone network using an event-driven network simulator (SuperMac).
This research is supported by France Télécom's research center, the Centre National d'études des Télécommunications (CNBT) under contract 94 1 B 003
This is a preview of subscription content, log in via an institution to check access.
Preview
Unable to display preview. Download preview PDF.
References
Bennani Y. & Bossaert F. (1996), “Predictive Neural Networks for Traffic Disturbance Detection in the Telephone Network”, Proceedings of the IEEE/IMACS CESA'96.
Bennani Y. (1992). “Connectionist Approach for Automatic Speaker Recognition: Modelling and Identification”, Ph.D. Thesis (in French), Order N° 1948 — LRI, University of Paris Sud, Orsay, France.
Bernani Y. (1993), “Probabilistic Cooperation Of Connectionist Expert Modules: Validation On A Speaker Identification Task”, IEEE, ICASSP'93, Minneapolis, Minnesota, U.S.A.
Bernani Y. (1995), “Modular and Hybrid Connectionist System for Automatic Speaker Identification”, Neural Computation, MIT Press, 7(4):791–797.
Boutleux E. & Dubuisson B. (1995), “Détection et Suivi d'évolutions de l'État d'un Système Complexe: Application au réseau Téléphonique Français”, Personal Communication.
Didelet E. (1994), “A Neural Technique Approach to Network Traffic Management”, ITC 14, Antibes, France.
Drucker H., Schapire R., Simard P. (1993), “Boosting Performance in Neural Networks”, International Journal of Pattern Recognition and Artificial Intelligence, 7(4):705–719.
Ghosh J., Beck S., Deuser L. (1996), “Integration of Neural Classifiers for Passive Sonar Signals”, In C.T. Leondes, editor, DSP Theory and Applications, Academic Press.
Guermeur Y & Gallinari P. (1996), “Combining Statistical Models for Protein Secondary Structure Prediction”, ICANN'96.
Hampshire J.B. & Waibel A.H. (1992), “The Meta-Pi Network: Building Distributed Representations for Robust Multisource Pattern Recognition” IEEE Transactions on Pattern Analysis and Machine Intelligence, 14(7):751–769.
Hansen L. K. & Salamon P. (1990), “Neural Network Ensembles”, Transactions on Pattern Analysis and Machine Intelligence, pp. 993–1001.
Herrmann F., Stern D., Chemouil P. (1989), “SUPERMAC: A Software Tool for the Performance Evaluation of Network Management”, ICCC Symp., Beijing, China.
Jacobs R.A., Jordan M.I., Nowlan S.J., Hinton G.E. (1991), “Adaptive Mixtures of Local Experts”, Neural Computation, 3:79–87.
Lamy B. & Fogelman F. (1995), “Can Multiple Initializations Improve Generalization ?”, ICANN'95, 1:39–44, Paris.
Leray P., Gallinari P., Didelet E. (1996), “Diagnosis Tools for Telecommunication Network Traffic Management”, ICANN'96.
Raczkiewicz M. & Stern D. (1993), “Methods to Detect Traffic Disturbances for Real-Time Network Management”, Technical Repport, 1993, DE/ATR/82-93.
Stern D. & Chemouil P. (1992), “A Diagnosis Expert System for Network Traffic Management”, Networks'92, Kobe, Japan.
Turner K. & Ghosh J. (1996), “Theoretical Foundations of Linear and Order Statistics Combiners for Neural Pattern Classifiers”, TR-95-02-98, CVRC, University of Texas, Austin.
Wolpert D. H. (1992), “Stacked Generalization”, Neural Computation, MIT Press, 5(2):241–259.
Xu L., Krzvzak A., Suen C. Y. (1992), “Methods of Combining Multiple Classifiers and Their Applications to Handwriting Recognition”, Transactions on Systems, Man and Cybernetics, 22(3):418–435.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1997 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Younès, B., Fabrice, B., Elisabeth, D. (1997). Linear and nonlinear combinations of connectionist models for local diagnosis in real-time telephone network traffic management. In: Gerstner, W., Germond, A., Hasler, M., Nicoud, JD. (eds) Artificial Neural Networks — ICANN'97. ICANN 1997. Lecture Notes in Computer Science, vol 1327. Springer, Berlin, Heidelberg. https://doi.org/10.1007/BFb0020298
Download citation
DOI: https://doi.org/10.1007/BFb0020298
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-63631-1
Online ISBN: 978-3-540-69620-9
eBook Packages: Springer Book Archive