A design of granular fuzzy classifier
作者：
Highlights：
• A new design methodology of granular fuzzy classifiers based on information granules is proposed.
• The information granules are constructed in terms of a geometrical distribution of patterns.
• The elements in an information granule are heterogeneous with respect to their class distributions.
• We improve the class homogeneity of the originally constructed information granules by using Particle Swarm Optimization.
• We consider the center points of the information granules as prototypes of the prototype based classifiers.
摘要
•A new design methodology of granular fuzzy classifiers based on information granules is proposed.•The information granules are constructed in terms of a geometrical distribution of patterns.•The elements in an information granule are heterogeneous with respect to their class distributions.•We improve the class homogeneity of the originally constructed information granules by using Particle Swarm Optimization.•We consider the center points of the information granules as prototypes of the prototype based classifiers.
论文关键词：Granular fuzzy classifier,Pattern classifier,Information granules,Weighting scheme
论文评审过程：Available online 9 May 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.04.040
