|Article title||EXTRACTION OF ATTRIBUTES FROM GRAPHS USING THE MECHANISTIC ANALOGY FUNCTIONALS FOR EXPERT SYSTEMS|
|Authors||A.N. Tselykh, V.S. Vasilev, L.A. Tselykh|
|Section||SECTION IV. DATA ANALYSIS AND KNOWLEDGE MANAGEMENT|
|Month, Year||08, 2016 @en|
|Abstract||In this paper we present a new approach to the development of the knowledge base of pro-duction Expert System (ES) based on the properties of cause-and-effect relations between the attrib-utes of a logical inference. It is proposed to use fuzzy cognitive maps, which are represented by fuzzy weighted signed directed graphs with feedback cycles; the graph stands as a base to generate antecedents and consequents set of the fuzzy rules to the ES based on the fuzzy inference engine. The resulting knowledge base uses the system qualities to improve the intelligence of logical inference. Modern graphs have a high degree of dimensionality and multiple cycles, which greatly complicate the heuristic definition of attributes of logical inference. The problem of correction of the feedback gear ratio is resolved by reducing the adjacency matrix (as it possible) to the upper (or lower) triangular form using the quadratic functionals of the mechanistic analogy (barrier penalty, "inverse"). The first type is a fixed signed convex functionals, so the minimization process has a unique solution obtained by any methods of the 0-th and 1-th order. Functionals of the second type are unbounded, but, as it is constantly mapping a real-valued solution to integer permutations, so the problem has also a solution that is obtained with the same methods. Minimization algorithm of functional presented by the paired comparisons method is presented. It is shown that the algorithm is computationally efficient, relevant to O(n3) time hardness. We conducted a series of numerical experiments to analyze and compare the quality of the solutions, as well as determine the best functional. The experiment showed efficiency for the matrix size of 103х103. The algorithm would require to solve the problem of accumulation of rounding errors in floating point for very large matrices (big data). Evaluation of the seven types of mechanistic analogies functional based on feasibility, performance, effectiveness and applicability of the subject area has been implemented. The approach of a sequence transformation of the original matrix of fuzzy graph model with two types of qualities functional is described. It allows generating the attributes of production ES.|
|Keywords||Expert systems; signed directed weighted cyclic graphs; functionals based on a mechanistic analogy.|
|References||1. Gaines B.R., Shaw M.L.G. Eliciting Knowledge and Transferring it Effectively to a Knowledge-Based System, In: IEEE Trans. Knowl. Data Eng. IEEE Press, New York, 1993, No. 5 (1), pp. 4-14
2. Nasuti F.W. Knowledge Acquisition Using Multiple Domain Experts in the Design and De-velopment of an Expert System for Disaster Recovery Planning. Doctoral Thesis Proposal. Nova Southeastern University, 2000, 206 p.
3. Feigenbaum E.A. The Art of Artificial Intelligence: Themes and Case Studies of Knowledge Engineering, In Proceedings Fifth International Joint Conference on Artificial Intelligence. Morgan Kaufmann, Los Altos, CA, 1977, pp. 1014-1029.
4. Hayes-Roth F., Waterman D.A.D., Lenat B. Building Expert Systems. London: Addison-Wesley, 1983, 444 p.
5. Bing W., Chenyan Z. Dynamics of Knowledge Acquisition via E-Learning Community. JCIT 8, 2013, pp. 168-175.
6. Kadhim M.A., Alam M.A., Kaur H. A Multi-intelligent Agent Architecture for Knowledge Ex-traction: Novel Approaches for Automatic Production Rules Extraction, J. IJMUE, 2014,
No. 9 (2), pp. 95-114.
7. Henry N., Fekete J.-D. Matrix Explorer: A Dual Representation System to Explore Social Networks, IEEE Trans Vis Comput graph, 2006, No. 12 (5), pp. 677-684.
8. Kormen T., Leyzerson Ch., Rivest Ronal'd, Shtayn K. Algoritmy, postroenie i analiz [Algo-rithms, construction and analysis]. Moscow: Vil'yams, 2005, 1296 p.
9. Gorbatov V.A., Smirnov M.I., Khlytchiev I.S. Logicheskoe upravlenie raspredelennymi siste-mami [Logical control of distributed systems]. Moscow: Energoatomizdat, 1991, 287 p.
10. Kofman A. Vvedenie v teoriyu nechetkikh mnozhestv [Introduction to the fuzzy sets theory]. Moscow: Radio i svyaz', 1982, 432 p.
11. Newman M.E.J. Finding community structure in networks using eigenvectors of matrices, Physical Review E, 2006, No. 74 (3).
12. Blondel V.D., Guillaume J.-L., Lambiotte R., Lefebvre E. Fast unfolding of communities in large networks, J. Stat. Mech. Theory and Experiment, 2008, No. 10.
13. Pons P., Latapy M. Computing communities in large networks using random walks, JGAA, 2006, No. 10 (2), pp. 191-218.
14. Ye Y., Jiang Z., Diao X., Du G. Extended event-condition-action rules and fuzzy Petri nets based exception handling for workflow management, Expert Syst. Appl., 2011, No. 38 (9),
15. Girvan V., Newman M.E.J. Community structure in social and biological networks, Proceed-ings of the National Academy of Sciences USA, 2002, No. 99 (12), pp. 7821-7826.
16. Zhang L., Du W., Song J.-J., Xiang H.-Y. Traffic assignment method based on dynamic fuzzy control and fuzzy decision making. Jiaotong Yunshu Gongcheng Xuebao, J. Traffic Trans-portation Engineering, 2010, No. 10 (3), pp. 110-117.
17. Kristianto Y., Gunasekaran A., Helo P., Hao Y. A model of resilient supply chain network de-sign: A two-stage programming with fuzzy shortest path, Expert Syst. Appl., 2014, No. 41 (1), pp. 39-49.
18. Radhavan U.N., Albert R., Kumara S. Near linear time algorithm to detect community structures in large-scale networks, Physical Review E, 2007, No. 76 (3).
19. Goyal S., Grover S. A fuzzy multi attribute decision making approach for evaluating effec-tiveness of advanced manufacturing technology - in Indian context, IJPQM, 2013, No. 11 (2), pp. 150-178.
20. Hipel K.W., Kilgour D.M., Bashar A.M. Fuzzy preferences in multiple participant decision making, J. Scientia Iranica, 2011, No. 18 (3), pp. 627-638.
21. Tselykh A., Vasilev V., Tselykh L. Fuzzy graphs clustering with quality relations functionals in cognitive models, Proceedings of the First International Scientific Conference “Intelligent In-formation Technologies for Industry” (IITI’16); series “Advances in Intelligent Systems and Computing”, Series Vol. 450, 2016, Vol. 1, pp. 349-360.