IMPLEMENTATION OF KEA-KEYPHRASE EXTRAC-TION ALGORITHM BY USING BISECTING K-MEANS CLUSTERING TECHNIQUE FOR LARGE AND DYNAMIC DATA SET

In most traditional techniques of document clustering, the number of total clusters is not known in advance and the cluster that contains the target information cannot be deter-mined since the semantic nature is not associated with the cluster. To solve this problem, this work proposes a new clustering algorithm based on the Kea[1] key phrase extrac-tion algorithm which returns several key phrases from the source documents by using some machine learning tech-niques. In this work, documents are grouped into several clusters like Bisecting K-means, but the number of clusters is automatically determined by the algorithm with some heu-ristics using the extracted key phrases. By this it is easy to extract test documents from massive quantities of resources.