Text clustering using fuzzy neighborhood and evaluation of clusters

This study is concerned with text clustering and evaluating the clusters. Fuzzy neighborhood, a method of text mining, is used for this purpose. Five different clustering algorithms are used, they are kernel affinity propagation, kernel hard c-means, kernel fuzzy c-means, kernel hard k-means++ and kernel variable size hard c-means. These algorithms are applied to the clustering of nouns and adjectives in Chinese documents and SNS. The results from the previous four algorithms except kernel affinity propagation are evaluated by Rand index.