An Efficient and Applicable Clustering Algorithm using Fuzzy ART

Lots of clustering algorithms have been developed, and in most of them some parameters should be determined by hand. However, it is very difficult to determine them manually without any prior domain knowledge. To solve this problem, a novel similarity based clustering algorithm was presented. It aimed at avoiding instructional parameters to be determined by hand, and at the same time, improving the efficiency of clustering. By introducing fuzzy Adaptive Resonance Theory (ART) to pre-train the similarity criterion, the instructional parameter was determined dynamically. The new clustering algorithm was analyzed and applied later to cluster the product data. A comparison with K-means shows that this algorithm produced clusters with higher quality and in less time. Validity analysis to the clustering results confirms applicability of this algorithm.