Evaluation of Text Clustering Based on Iterative Classification

Text clustering is a useful and inexpensive way to organize vast text repositories into meaningful topics categories. Although text clustering can be seen as an alternative to supervised text categorization, the question remains of how to determine if the resulting clusters are of sufficient quality in a real-life application. However, it is difficult to evaluate a given clustering of documents. Furthermore, the existing quality measures rely on the labor standard, which is difficult and time-consuming. The need for fair methods that can assess the validation of clustering results is becoming more and more critical. In this paper, we propose and experiment an innovative evaluation measure that allows one to effectively and correctly assess the clustering results.