The IEMA Fuzzy c-Means Algorithm for Text Clustering

The fuzzy c-means algorithm is a soft version of the popular k-means clustering. As is well known, the k-means method begins with an initial set of randomly selected exemplars and iteratively refines this set so as to decrease the sum of squared errors. The k-centers clustering is moderately sensitive to the initial selection of centers, so it is usually rerun many times with different initializations in an attempt to find a good solution. We propose a new version of the fuzzy c-means algorithm for unstructured data to detect the best centroid of clusters, and we choose the final partition according to the validation of the Xie-Beni index. We apply our method to three different corpora (literature, forum, and ads) to verify the quality of the procedures.