Words Clustering Based on Keywords Indexing from Large-scale Categorization Corpora
Liu Hua · 2009
Keywords are indexed automatically for large-scale categorization corpora. Indexed keywords of more than 20 documents are selected as seed words, thus overcoming subjectivity of selecting seed words in clustering; at the same time, clustering is limited to particular category corpora and keywords indexed feature extraction method is adopted to obtain domanial words automatically, thus reducing noise of similarity calculation.