ABDUCTION APPLICATION IN PROBLEMS OF CLASSIFICATION OF DATA ABOUT OIL AND GAS FACILITIES
Keywords:
optimization, intelligent decision support, drilling of oil and gas wells, objective functions, rules, knowledge base, abductive framework, confidence coefficient, limits.Abstract
The research is devoted to utilization of the abductive reasoning means for data extraction problems. The conducted research shows that data classification can be interpreted as one of the abductive logic programming problems, which allows utilizing of user-defined domain restrictions. Interpretation of classification models based on decision trees made in accordance with the abductive method using domain restrictions allows increasing the efficiency in the case of partial lack of input data. In order to consider the probabilistic information with the help of the basic and output formal theories, the overall framework was also extended to abductive framework that is based on cost factors and can be used for data mining applications, which will ultimately improve the overall quality of the results. Thus, it was shown that abductive reasoning can be used in the context of classification problems to explain the course of reasoning of the made classification and improve the overall efficiency in the event of the system operation with the partial absence of the input data and external domain knowledge. This approach can be improved by combining different data mining paradigms such as classification, clustering, association rules and by utilizing abductive framework with restrictions.
Downloads
References
2 Liu, B.; Hsu, W. & Ma, Y., Integrating classification and association rule mining. Proceedings of the 4th international conference on Knowledge Discovery and Data mining (KDD'98), AAAI Press, 1998, pp. 80-86.
3 Wenmin Li. CMAR: Accurate and efficient classification based on multiple classassociation rules / Wenmin Li, Jiawei Han, and Jian Pei. – In ICDM, 2001. – Р. 369-376.
4 Вагин В.Н. Абдукция в задачах планирования работ в сложных объектах / В.Н. Вагин, К.Ю. Хотимчук // Искусственный интеллект и принятие решений. – М.: Ленанд, 2011. – Т.1. – С. 3-13.
5 Quinlan J.R. Improved use of continuous attributes in C4.5 / Quinlan J.R. // Journal of Artificial Intelligence Research. – 1996. – V.4. – P. 77-90.
6 Mancarella P. An abductive proof procedure handling active rules / P. Mancarella and G. Terreni // AI*IA 2003: Advances in Artificial Intelligence / A.Cappelli and F.Turini, editors. – SV of LNAI, 2003.– Р. 105-117.
7 Mohammed J. Zaki. Theoretical foundations of association rules / Mohammed J. Zaki and Mitsunori Ogihara; in Proceedings of 3 rd SIGMOD’98 Workshop on Research Issues in Data Mining and Knowledge Discovery (DMKD’98), Seattle, Washington, 1998. – P. 10- 17.
8 Eskilson, J. & Carlsson, M., SICStus MT - Multithreaded Execution Environment for SICStus Prolog. Konstantinos F. Sagonas, ed. Implementation Technology for Programming Languages based on Logic, 1998, pp. 59-71.
9 Kakas A.C. Abductive concept learning / Antonis C. Kakas, F. Riguzzi // New Generation Computing, 2000. – V. 18(3). – Р. 243-294.
10 Michell T. Machine Learning / T. Michell. – McGraw Hill, 1997. – 414 p.
11 Kakas A.C. Aclp: Abductive constraint logic programming / A.C. Kakas, A. Michael, and C. Mourlas // Journal of Logic Programming, 2000. – V. 44 (1-3). – Р. 129-177.
12 Dubois D. Automated reasoning using possibilistic logic: Semantics, belief revision, and variable certainty weights / D. Dubois, J. Lang, H Prade. – Knowledge and Data Engineering. – 1994. – Feb. – P. 64-71.
.png)
1.png){kind=logo}
