A Validity Measure for Fuzzy Clustering
IEEE Transactions on Pattern Analysis and Machine Intelligence
Normalized Cuts and Image Segmentation
IEEE Transactions on Pattern Analysis and Machine Intelligence
The Structure and Dynamics of Networks: (Princeton Studies in Complexity)
The Structure and Dynamics of Networks: (Princeton Studies in Complexity)
Comparative analysis for k-means algorithms in network community detection
ISICA'10 Proceedings of the 5th international conference on Advances in computation and intelligence
| Hi-index | 0.00 |
Fuzzy cluster validity criterion tends to evaluate the quality of fuzzy partitions produced by fuzzy clustering algorithms. In this paper, an effective validity index for network fuzzy clustering is proposed, which involves the compactness and separation measures for each cluster. The simulated annealing strategy is used to minimize this validity index, associating with a dissimilarity-index-based fuzzy c-means iterative procedure, under the framework of a random walker Markovian dynamics on the network. The proposed algorithm (SADIF) can efficiently identify the probabilities of each node belonging to different clusters during the cooling process. An appropriate number of clusters can be automatically determined without any prior knowledge about the network structure. The computational results on several artificial and real-world networks confirm the capability of the algorithm.