Logical Reduction of Tabular Systems

• Autorzy: Antoni Ligęza
• Języki publikacji: EN

Abstrakty

EN

Reduction of various sets of data and knowledge bases constitutes an important operation, possibly leading to generation of some minimal or at least reduced form. This is important both for design and analysis of knowledge bases, as well as for enabling perception and better comprehension of large amounts of data. This paper presents an algorithmic approach and practical results concerning reduction of tabular systems. Each table of such system has a specific number of columns labeled with selected attributes. The rows contain attribute values referring to a specific object. Reduction consists in gluing rows of the table. Any table of a tabular system can contain description of certain objects or specification of rules. The main issue concerning tabular systems presented and discussed in this paper is reduction (or minimization) of such systems, i.e. the problem consisting in finding a possibly most concise representation of the data or knowledge. This kind of operation can be referred to as system reduction or finding a minimal form, but one logically equivalent to the input one. Reduction of a tabular system leads to granular knowledge specification. In case of pure data one obtains the so-called data templates. The data templates forms a granular representation comprising many atomic records, expressed with non-atomic values, such as intervals, sets, or structured objects. In case of reduction of tabular systems, rules with the same conclusion are "glued" together; the reduction operation is applied in fact to the precondition part of the rules. The paper presents the idea of tabular systems and logical foundations of reduction with use of backward dual resolution. Two algorithms of reduction are outlined. Further, example results of reduction of some tabular systems are presented. (original abstract)

W niniejszej publikacji zaprezentowano algorytmiczne podejście i praktyczne wyniki oparte na redukcji systemów tablicowych. Każda tabela takiego systemu ma specyficzną liczbę kolumn oznaczoną wybranymi atrybutami. Wiersze tabeli zawierają atrybuty wartości odnoszące się do wybranego obiektu. Redukcja polega na sklejaniu wierszy tabeli. Każda tabela systemu tablicowego może zawierać opis wybranych obiektów lub specyfikacji zasad. Głównym zagadnieniem związanym z systemami tablicowymi zaprezentowanym i omówionym w tym artykule jest redukcja (lub minimalizacja) tych systemów. (AŁ)

Słowa kluczowe

EN

Knowledge based systems; Blackboard architecture; Data processing

PL

Systemy z bazą wiedzy; Architektura tablicowa; Przetwarzanie danych

• Czasopismo: Prace Naukowe Akademii Ekonomicznej we Wrocławiu
• Rocznik: 2004
• Numer: nr 1011 Pozyskiwanie wiedzy i zarządzanie wiedzą
• Strony: 188--199

Twórcy

autor

Antoni Ligęza

Bibliografia

• Bancilhon F., Khoshafian S.: A calculus for complex objects, .Journal of Computer and Systems Sciences", 38,1989, 326-340.
• Coenen, F. et al.: Validation and verification of knowledge based systems: Report on EUROVAV'99. "Knowledge Engineering Review" Vol. 15:2,2000,187-196.
• Conolly Т., Begg C, Strachan A.: Database Systems. A Practical Approach to Design, Implementation and Management, Harlow, England, Addison-Wesley, 1999.
• Gatnar E: Symboliczne metody klasyfikacji danych. Warszawa, Wydawnictwo Naukowe PWN 1998.
• Ligęza A.: Elements of algebraic data analysis for verification of qualitative properties. In: A. Baborski (Ed.), Pozyskiwanie wiedzy z baz danych, Wyd. Akademii Ekonomicznej we Wrocławiu, Turawa 2000, 18-28.
• Ligęza A.: Logical analysis of databases and rule-based systems. In: R. Trappl (Ed.), Cybernetics and Systems 2000, Austrian Society for Cybernetic Studies, Vienna, 2000, 757-762.
• Ligęza A. (200la): Towards logical analysis of tabular rule-based systems, "International Journal of Intelligent Systems", 16, 333-360.
• Ligęza A. (200 Ib): Knowledge abstraction and algebra of tabular systems. In: A. Baborski (Ed.) Pozyskiwanie wiedzy z baz danych, Karpacz, 2001.
• Liu B., Ku L.P., Hsu W.: Discovering interesting holes in data. "Proceedings of IJCAI'87" Nagoya Vol. 2 1997 930-935.
• Pawlak Z.: Rough Sets. Theoretical Aspects of Reasoning about Data. Dordrecht, Kluwer Academic Publishers 1991.
• Rokita Ł.: Selected methods for analysis of tabular rule-based systems. M.Sc. Thesis, UST- AGH, Cracow, 2003 (in Polish).