A connectionist indexing approach for CBR systems
An important factor that plays a major role in determining the performances of a CBR system is the complexity and the accuracy of the case retrieval phase. Both flat memory and inductive approaches suffer from serious drawbacks. In the first approach, the search time becomes considerable when dealing with large scale memory base, while in the second one the modifications of the case memory becomes very complex because of its sophisticated architecture.
In this paper, we show how we can construct a simple efficient indexing system structure. We construct a case hierarchy with two levels of memory: the lower level contains cases organised into groups of similar cases, while the upper level contains prototypes, each of which represents one group of cases. The construction of prototypes is made by using an incremental prototype-based network. This upper level parallel memory is used as an indexing system during the retrieval phase.
Key wordsIndexing Incremental Neural Network Prototype
Unable to display preview. Download preview PDF.
- 1.A. Aamodt and E. Plaza. Case-based reasoning: Foundational issues, methodological variations, and system approaches. AICOM, 7(1), March 1994.Google Scholar
- 2.E. Alpaydin. Gal: Networks that grow when they learn and shrink when they forget. Technical Report TR 91-032, International Computer Science Institute, May 1991.Google Scholar
- 3.A. Azcarraza and A. Giacometti. A prototype-based incremental network model for classification task. In Fourth International Conference on Neural Networks and Their Applications, Nimes, France, 1991.Google Scholar
- 4.S.K. Bamberger and K. Goos. Integration of case-based reasoning and inductive learning methods. In First European Workshop on CBR, November 1993.Google Scholar
- 5.R. Barletta. An introduction to case-based reasoning. AI Expert, August 1991.Google Scholar
- 6.A. Giacometti. Modèles Hybrides de l'Expertise. PhD thesis, Telecom-Paris, 1992.Google Scholar
- 7.J. Kolodner. Case-Based Reasoning. Morgan Kaufmann Publishers, Inc, 1993.Google Scholar
- 8.M. Malek and V. Rialle. A case-based reasoning system applied to neuropathy diagnosis. In Second European Workshop, EWCBR-94 Lecture Notes in Computer Science. Springer-Verlag, November 1994.Google Scholar
- 9.M. Manago, K. Althoff, E. Auriol, R. Traphoner, S. Wess, N. Conruyt, and F. Maurer. Induction and reasoning from cases. In First European Workshop on CBR, November 1993.Google Scholar
- 10.P. Myllymaki and H.Tirri. Massively parallel case-based reasoning with probabilistic similarity metrics. In First European Workshop on CBR, November 1993.Google Scholar
- 11.J.R. Quinlan. Induction of decision trees. Machine Learning, (1):81–106, 1986.Google Scholar
- 12.J.R. Quinlan. C4.5. Morgan Kaufmann Publishers, 1992.Google Scholar
- 13.D.L. Reilly, L.N. Cooper, and C. Elbaum. A neural model for category learning. Biological Cybermetics, (45):35–41, 1982.Google Scholar
- 14.C.K. Riesbeck and R.C. Schank. Inside Case-Based Reasoning. Lawrence Erlbaum Associates, publishers, 1989.Google Scholar
- 15.P. Thrift. A neural network model for case-based reasoning. In Proceddings of the DARPA Case-Based Reasoning Workshop, May 1989.Google Scholar
- 16.S.B. Thrun, J. Bala, E. Bloedorn, I. Bratko, B. Cestnik, J. Cheng, K. De Jong, S Dzroski, S.E. Fahlman, D. Fisher, R. Hamann, K. Kaufman, S. Keller, I. Kononenko, J. Kreuziger, R.S. Michalski, T. Mitchell, P. Pachowicz, Y. Reich H. Vafaie, W. Van de Welde, W. Wenzel, J. Wnek, and J. Zhang. The monk's problems a performance comparison of different learning algorihms. Technical Report CMU-CS-91-197, Carnegie Mellon University, December 1991.Google Scholar
- 17.P.E. Utgoff. Incremental induction of decision trees. Machine Learning, (4):161–186, 1989.Google Scholar