Case-based reasoning (CBR) is an approach to problem solving by reuse of previously made experience.
The basic idea of CBR is to solve new problems by applying solutions of similar past problems. In order to realize this idea a CBR-solver is required to incorporate several approaches, among them a retrieval approach used for finding the similar problems and their solutions as well as a reuse approach used for adaptation of found solutions to fit the new problem.
The retrieval approach proposed by CBR is proven useful also for scenarios which don’t require the problem solving, but only a search technology. It has been successfully integrated in many applications, for example in applications within electronic commerce. Its advantage over the traditional database search is the ability of finding useful objects, such as consumer products or wiki pages, which don’t match the query exactly. A result list contains found objects arranged in descending order of similarity to the query. The similarity measure plays a central role in case-based reasoning since it allows a comparison of a new problem (a query) with previously solved problems (stored objects). It’s usually modelled by a domain expert, who represents on this way her/his knowledge of similarity in the respective domain.
While the basic idea and underlying theories were developed in the US, now CBR is intensively researched in Europe and other countries as well. Meanwhile there are diverse domain-specific and domain-independent, free and commercial systems available, which successfully apply the concepts of CBR.
However, the CBR technology doesn’t support sufficiently special “complex” objects, called generalized cases, occurring in problem solving scenarios as well as in scenarios requiring only search. Contrary to trivial objects which can be represented by use of feature vectors, a representation of “complex” objects requires variables and constraints defining dependencies between them. For example such objects occur in the domain of electronic commerce where they are descriptions of parameterized and configurable products.
The goal of GenCase is to develop a unified view on “complex” objects and a methodology for similarty-based reasoning with them, in particular, methods for representation, similarity assessment, as well as index-based retrieval.
Following members of the Wi2-group are working or have been working within GenCase:
The project GenCase is funded by the University of Trier, Germany
|•||Alexander Tartakovski, Martin Schaaf, and Ralph Bergmann. Retrieval and Configuration of Life Insurance Policies. In Hector Munoz-Avila, and F. Ricci, editors, Sixth International Conference on Case-Based Reasoning (ICCBR 2005), volume 3620 , pages 552-565, Chicago, Illinois (USA), August 2005, Springer.|
|•||Alexander Tartakovski, Martin Schaaf, and Rainer Maximini. Applying Generalized Cases to Retrieval and Configuration of Life Insurance Policies. In Klaus-Dieter Althoff, Andreas Dengel, Ralph Bergmann, Markus Nick, and Thomas Roth-Berghofer, editors, Professional Knowledge Management: Third Biennial Conference, WM 2005, Kaiserslautern, Germany, April 10-13, 2005, Revised Selected Papers, volume 3782 of LNAI, pages 293 - 303, December 2005, Springer-Verlag GmbH.|
|•||Alexander Tartakovski, Martin Schaaf, and Rainer Maximini. Optimization Based Retrieval and Configuration of Temporary Life Insurance Policies. In Klaus-Dieter Althoff, Andreas Dengel, Ralph Bergmann, Markus Nick, and Thomas Roth-Berghofer, editors, WM2005: Professional Knowledge Management, Experience and Visions, pages 269-274, Kaiserslautern, April 2005, German Research Center for Artificial Intelligence DFKI GmbH.|
|•||Alexander Tartakovski, Martin Schaaf, Rainer Maximini, and Ralph Bergmann. MINLP Based Retrieval of Generalized Cases, Proceedings of the 7th European Conference on Case Based Reasoning, ECCBR 2004. In Peter Funk, and Pedro A. González Calero, editors, Advances in Case-Based Reasoning, volume 3155 of LNAI, pages 404-418, Madrid, Spain, August 2004, Springer Verlag, Berlin-Heidelberg.|
PDF © Springer
|•||Martin Schaaf, Andrea Freßmann, Rainer Maximini, Ralph Bergmann, Alexander Tartakovski, and Martin Radetzki. Intelligent IP Retrieval Driven by Application Requirements. Integration, the VLSI Journal, 37(4):253-287, 2004.|
PDF © ELSEVIER
|•||Rainer Maximini, Alexander Tartakovski, and Ralph Bergmann. Investigating different Methods for efficient Retrieval of Generalized Cases. In Ulrich Reimer, Andreas Abecker, Steffen Staab, and Gerd Stumme, editors, Professionelles Wissensmanagement -Erfahrungen und Visionen (WM 2003), pages 303-304, Luzern, Switzerland, April 2003, GI.|
PS © GI
|•||Alexander Tartakovski, and Rainer Maximini. Similarity Assessment and Retrieval of Generalized Cases. In Workshop on Knowledge and Experience Management (FGWM 2003), Karlsruhe Germany, October 2003.|
PDF © FGWM
|•||Rainer Maximini, and Alexander Tartakovski. Approximative Retrieval of Attribute Dependent Generalized Cases. In Workshop on Knowledge and Experience Management (FGWM 2003), Karlsruhe Germany, October 2003.|
PDF © FGWM
|•||Babak Mougouie, and Michael M. Richter. Generalized Cases, Similarity and Optimization. In D. Hutter, and W. Stephan, editors, Deduction and beyond, 2003, LNAI 2605, Springer-Verlag.|
|•||Babak Mougouie, Michael M. Richter, and Ralph Bergmann. Diversity-Conscious Retrieval from Generalized Cases: A Branch and Bound Algorithm. In Kevin D. Ashley, and Derek Bridge, editors, 5th International Conference on Case-Based Reasoning (ICCBR 2003), LNAI 2689, pages 319-331, 2003, Springer-Verlag.|
PS © Springer
|•||Babak Mougouie, and Ralph Bergmann. Similarity Assessment for Generalized Cases by Optimization Methods. In S. Craw, and A. Preece, editors, European Conference on Case-Based Reasoning (ECCBR'02), volume 2416 of LNAI, pages 249-263, 2002, Springer.|
PDF © Springer