Supervised Object Classification Using Adaptive Active Hypercontours with Growing Neural Gas Representation

• Autorzy: Pryczek M.
• Języki publikacji: EN

Abstrakty

EN

In this paper a new classifier induction method is proposed. It is based on Hebbian-like competitive learning idea and modified Adaptive Active Hypercontour algorithm- recently suggested as general purpose classifier induction technique. Adaptive neural networks utilising Growing Neural Gas algorithm are here used as reference implementation aiming to ease induction process in high-dimensional spaces. Presented work focuses on supervised classification problem: textual document classification and transformer malfunction recognition have been chosen as evaluation fields.

Słowa kluczowe

EN

neural network; pattern analysis; classification system

PL

sieć neuronowa; system klasyfikacji

• Wydawca: Wydawnictwo Politechniki Łódzkiej
• Czasopismo: Journal of Applied Computer Science
• Rocznik: 2008
• Tom: Vol. 16, nr 2
• Strony: 67--80
• Opis fizyczny: Bibliogr. 18 poz.

Twórcy

autor

Pryczek M.

• Technical University of Lodz Institute of Computer Science ul. Wolczanska 215, 90-924 Lodz, Poland,

Bibliografia

• [1] Caselles V., Kimmel R., Sapiro G.: Geodesic Active Contours, in: International Journal of Computer Vision, 22(1), 1997, pp. 61-79.
• [2] Fritzke B.: A growing neural gas network learns topologies, in: Tesauro G, Touretzky DS, Leen TK (eds) Neural Information Processing Systems, MIT Press, vol 7, 1994, pp. 625-632.
• [3] Haykin S.: Neural Networks: A Comprehensive Foundation, Macmillan College Publishing Company, New York, 1994.
• [4] International Electrotechnical Commission: Interpretation of the Analysis of Gases in Transformers and other Oil-Filled Electrical Equipment in Service. Geneva, 1979.
• [5] Kass M., Witkin W., Terzopoulos D.: Snakes: Active Contour Models, in: International Journal of Computer Vision, 1988, pp. 321-331.
• [6] Rudnicki M., Szczepaniak P.S., Cholajda P.: Genetic Algorithm as a Tool for Solving Electrical Engineering Problems. in: K.Miettinen, P.Neittaanmäki, and J.Périaux (Eds.), Evolutionary Algorithms in Engineering and Computer Science, Chapter 13, John Wiley & Sons, Chichester, 1999, ISBN 0 471 999 0 24, pp. 261- 281.
• [7] Szczepaniak P.S.: Fuzzy and Genetic Approach to Diagnosis of Power Transformers. in: Preprints of the 4th IFAC Symposium on Fault Detection, Supervision and Safety for Technical Processes, vol. 1, Budapest, Hungary, 2000, pp. 428-433.
• [8] Szczepaniak P.S.: Preprocessing of Chromatographic Data Reflecting State of Power Transformers. in: Proceedings of the FSKD'02 Conference, Singapore, 2002.
• [9] Szczepaniak P.S., Cholajda P.: Genetic interpretation of DGA results in power transformers. in: Proceedings. Int. Conf. SSCC'98: Advances in Systems, Signals, Control and Computers, Durban, South Africa, 1998, vol. II, pp. 216-220.
• [10] Szczepaniak P.S., Rudnicki M.: Soft-Computing Methods for Diagnosis and Design of Electrical Devices. in: Proceedings of IEEE Africon'99, Cape Town, South Africa, 1999, vol.2, pp. 753-758.
• [11] Szczepaniak P.S., Sasiak K.: Zastosowanie adaptacyjnych sieci logicznych do diagnostyki transformatorów. (Applications of adaptive logical networks to diagnosis of power transformers) (in Polish). in: Pomiary, automatyka, kontrola. 9(2005), ISSN 0032-4110, pp. 206-209.
• [12] Szczepaniak PS., Tomczyk A., Pryczek M.:Supervised web document classification using discrete transforms, active hypercontours and expert knowledge. in: Zhong N, Liu J, Yao Y, Wu JL, Lu S, Li K (eds): Web Intelligence Meets Brain Informatics. First WICI International Workshop, WImBI 2006, Beijing, China, 15-16 December 2006, Revised Selected and Invited Papers, LNCS 4845, LNAI, ISBN: 978-3-540-77027-5, Springer, 2007, pp. 305-323.
• [13] Tomczyk A. Active hypercontours and contextual classification. in: 5th International Conference on Inteligent Systems Design and Applications (ISDA), IEEE Computer Society Press, 2005, pp. 256-261.
• [14] Tomczyk A., Szczepaniak PS.: Adaptive potential active hypercontours, in: 8th International Conference on Artificial Intelligence and Soft Computing (ICAISC). Zakopane. Poland, Springer-Verlag, Berlin. Heidelberg, 2006, pp 692-701.
• [15] Tomczyk A., Szczepaniak P.S.: On the Relationship between Active Contours and Contextual Classification, in: 4th International Conference on Computer Recognition Systems (CORES), Poland, Springer, 2005, pp. 303-311.
• [16] Tomczyk A., Szczepaniak PS., Pryczek M.: Active contours as knowledge discovery methods. in: Discovery Science, Proceedings of 10th International Conference on Discovery Science (DS 2007), Sendai, Japan, October, 1-4 2007. LNCS 4755, LNAI, ISBN 978-3-540-75487-9. Springer, 2007.
• [17] Wiebe J., Wilson T., Cardie C.: Annotating expressions of opinions and emotions in language. in: Language Resources and Evaluation (formerly Computers and the Humanities), volume 39, issue 2-3, 2005, pp. 165-210.
• [18] Multi-Perspective Question Answering Corpus, release 1.2: http://www.cs.pitt.edu/mpqa