DOCUMENT CLASSIFICATION PROGRAM, SERVER, AND METHOD BASED ON OUTLINE FEATURE OF DOCUMENT INFORMATION

摘要

PROBLEM TO BE SOLVED: To provide a document classification program for classifying Web document information such as illegal and harmful sites at high speed without analyzing text content or image content in the document information.;SOLUTION: The document classification program makes a computer function as: an object featured value extraction means for extracting the object featured values of k dimensional vectors based on tag elements (m pieces of types) from object document information to be analyzed; a featured value determination means for determining whether the object featured values corresponding to the object document information belong to the prescribed range of learning featured values obtained from a plurality of document information included in a specific category; and a category classification means for classifying the object document information determined to be true by the featured value determination means as that included in the specific category.;COPYRIGHT: (C)2011,JPO&INPIT