Present work is a technique fuzzy text categorization followed by extractive summarization of categorized texts. At the onset, the texts of different subjects are fuzzy categorized based on relative matching with index terms of corresponding subjects. After forming the categorical groups, extractive summarization is performed on each text of each category. The fuzzy categorization is evaluated with fuzzy confusion matrix. The performance evaluation of this fuzzy categorization with Holdout method in terms of accuracy, precision, recall and fscore is appreciably high. The accuracy of summarization is evaluated using human generated summary and is fair. Also the categorization and summarization time is acceptable.
展开▼