Facebook Instagram Twitter RSS Feed PodBean Back to top on side

Learning strategies for neuro-fuzzy classification

In: Tatra Mountains Mathematical Publications, vol. 16, no. 2
Gabriela Andrejková - Henrich Tóth
Detaily:
Rok, strany: 1999, 211 - 224
O článku:
Neuro-fuzzy classification systems (NFCS) contain the fuzzy classification rules which are usually obtained by some algorithm. If the algorithm uses the data for the system construction then the algorithm is called a learning algorithm. D. Nauck and R. Kruse proposed some learning strategies for NFCS in [D. Nauck, R. Kruse: NEFCLASS–a neuro-ruzzy approach for the classification of data, Applied Computing 1995, Proc. of the ACM Symposium od Applied Computing, Nashville, Feb. 26–28, ACM Press, New York, 1995, pp. 461–65], [D. Nauck, R. Kruse: New Learning Strategies for NEFCLASS, in: Proceedings of 7-th IFSA World Congress, Prague, June 25–29, 1997, pp. 50–55]. In this paper we discuss the learning algorithms for NFCS based on the modification of the fuzzy partitions defined on each single dimension and on the sequential improving of the knowledge base. The goal is to derive fuzzy rules from a data set that can be separated in different crisp classes. Our applications are concerned with using NFCS for the prediction of Geomagnetic Storms [G. Andrejková, H. Tóth, K. Kudela: Fuzzy Neural Networks in the Prediction of Geomagnetic Storms, in: Proceedings “Artificial Intelligence in STP ”, Lund, 1997, pp. 173–179].
Ako citovať:
ISO 690:
Andrejková, G., Tóth, H. 1999. Learning strategies for neuro-fuzzy classification. In Tatra Mountains Mathematical Publications, vol. 16, no.2, pp. 211-224. 1210-3195.

APA:
Andrejková, G., Tóth, H. (1999). Learning strategies for neuro-fuzzy classification. Tatra Mountains Mathematical Publications, 16(2), 211-224. 1210-3195.