Bagged Nonlinear Hebbian Learning Algorithm for Fuzzy Cognitive Maps Working on Classification Tasks
Learning of fuzzy cognitive maps (FCMs) is one of the most useful characteristics which have a high impact on modeling and inference capabilities of them. The learning approaches for FCMs are concentrated on learning the connection matrix, based either on expert intervention and/or on the available historical data. Most learning approaches for FCMs are Hebbian-based and evolutionary-based algorithms. A new learning algorithm for FCMs is proposed in this research work, inheriting the main aspects of the bagging approach which is an ensemble based learning approach. The FCM nonlinear Hebbian learning (NHL) algorithm enhanced by the bagging technique is investigated contributing to an approach where the model is trained using NHL algorithm as a base learner classifier. This work is inspired from the neural networks ensembles and it is used to learn the FCM ensembles produced by the NHL exploiting better classification accuracies.
KeywordsEnsemble Method Autistic Disorder Ensemble Learning Connection Matrix Neural Network Ensemble
Unable to display preview. Download preview PDF.
- 3.Papageorgiou, E.I.: A Review Study of FCMs Applications during the last decade. In: Porc. FUZZ-IEEE 2011, Taipei, Taiwan, June 27-30, pp. 828–835 (2011)Google Scholar
- 4.Papageorgiou, E.I.: Learning Algorithms for Fuzzy Cognitive Maps: A Review Study. IEEE Transactions on SMC Part C (2011) (in press)Google Scholar
- 6.Froelich, W., Wakulicz-Deja, A.: Mining temporal medical data using adaptive fuzzy cognitive maps. In: Proc. 2nd Conf. on Human System Interactions, HSI 2009, art. no. 5090946, pp. 16–23 (2009)Google Scholar
- 12.Policar, R.: Ensemble based systems in decision making, IEEE Circuits and Systems Magazine, third quarter, 21–46 (2006) Google Scholar
- 13.Zhou, Z.-H.: Ensemble learning. Encyclopedia of Biometrics, 270–273 (2009)Google Scholar
- 14.Dietterich, T.G.: Machine learning research: Four current directions. AI Magazine 18(4), 97–136 (1997)Google Scholar