Modular Neuro-Fuzzy Networks Used in Explicit and Implicit Knowledge Integration
Proceedings of the Fifteenth International Florida Artificial Intelligence Research Society Conference
A framework for simultaneous co-clustering and learning from complex data
Proceedings of the 13th ACM SIGKDD international conference on Knowledge discovery and data mining
SCOAL: A framework for simultaneous co-clustering and learning from complex data
ACM Transactions on Knowledge Discovery from Data (TKDD)
| Hi-index | 0.00 |
This paper proposes a new approach to further refine and validate clusters using a multi-label voting algorithm to identify and classify similar objects by neuro-fuzzy classifier ensembles. The algorithm uses predictions of neuro-fuzzy experts trained on provisional clusters of heterogeneous collections of data. The multi-label predictions of the modular ensemble of classifiers are further combined, using fuzzy aggregation techniques. The proposed refinement algorithm considers then the votes, triggered by the confirmation of the classifiers' expertise for voted labels, and updates the clustering solution. Experiments on a Visual Arts objects database of color features show better interpretations and performances of the clusters inferred by the proposed algorithm. Its results can be widely used in various classification and clustering applications.