Tharam S. Dillon, Simon C.K. Shiu, Sankar K. Pal
2004 Applied intelligence (Boston)  
Preface There has recently been a spurt of activity to integrate different computing paradigms such as fuzzy set theory, neural networks, genetic algorithms, and rough set theory, for generating more efficient hybrid systems that can be classified as soft computing methodologies. Here the individual tool acts synergistically, not competitively, for enhancing the application domain of each other. The purpose is to provide flexible information processing systems that can exploit the tolerance for
more » ... imprecision, uncertainty, approximate reasoning, and partial truth in order to achieve tractability, robustness, low solution cost and close resemblance with human like decision making. Neuro-fuzzy computing, capturing the merits of fuzzy set theory and artificial neural networks, constitutes one of the best-known visible hybridizations encompassed in soft computing. This integration promises to provide, to a great extent, more intelligent systems (in terms of parallelism, fault tolerance, adaptivity, and uncertainty management) to handle real life ambiguous recognition/decision-making problems. Case based reasoning (CBR) may be defined as a model of reasoning that incorporates problem solving, understanding and learning, and integrates all of them with memory processes. It involves adapting old solutions to meet new demands, using old cases to explain new situations or to justify new solutions, and reasoning from precedents to interpret a new situation. Cases are nothing but some typical situations, already experienced by the system. A case may be defined as a contextualized piece of knowledge representing an experience that teaches a lesson fundamental to achieving the goals of the system. The system learns as a by-product of its reasoning activity. It becomes more efficient and more competent as a result of storing the experience of the system and referring to them in later reasoning. Systems based on this concept are finding widespread applications in problems like medical diagnosis and law interpretation where the knowledge available is incomplete and/or evidence is sparse. Research integrating CBR with a soft computing framework for developing efficient methodologies and algorithms for various decision making applications has drawn the attention of scientists. Here soft computing tools become effective in tasks like extracting cases from ambiguous situations and handling uncertainties, adapting new cases, retrieving old cases, discarding faulty cases, finding similarity between cases, maintaining an optimal size of case bases, and in approximating reasoning for justifying a decision. Design of efficient knowledge-based networks in the said paradigm is also being attempted. At present, the results on these investigations, both theory and applications, are being available in different journals and conference proceedings mainly in the fields of computer science, information technology, engineering and mathematics. The objective of this special issue is to present a cross sectional view of the present status of the said research demonstrating the role of soft computing tools, both individually and in combination, for performing different tasks of CBR with real life applications. Out of the twenty-three submissions, six papers were finally selected which address various methodologies, systems and applications. These are authored by experts from different active groups in the USA, Canada, France, India, Hong Kong, Spain and each of them is reviewed by two to three referees. The issue begins with a title article written by two editors, Pal and Shiu. It explains the concepts and features of CBR along with the relevance of soft computing. This introductory paper will enable the readers to understand the remaining articles better. A simultaneous optimization method of a CBR system using a genetic algorithm (GA) for financial forecasting is then explained by Kim. In this study, the GA simultaneously optimizes multiple factors, such as case indexing and representation, of the CBR system. Comparison of the GA approach with other conventional approaches for financial forecasting has also been carried out. The experimental result demonstrates the optimization ability of using GA for building CBR systems.
doi:10.1023/ fatcat:qdl3bjsls5f4bdtykj4l5gkz5a