This paper proposes a family of weighted fuzzy learning vector quantization algorithms, which include as a special case the existing fuzzy learning vector quantization algorithms. Under certain conditions, the proposed algorithms result in clustering algorithms that can also be derived using alternating optimization. The original fuzzy c-means (FCM) and generalized FCM (GFCM) algorithms can be obtained as a special case of the resulting clustering algorithms. The proposed formulation also provides the basis for the development of weighted GFCM algorithms, which are experimentally evaluated and compared with existing clustering algorithms.
展开▼
