SIMPLE SENTENCE CLASSIFYING SYSTEM BY SEMANTIC CONTENTS

摘要

PURPOSE:To execute the input sentence analysis processing in a short time, and to economically realize the input sentence analysis use knowledge construction, by grouping automatically a sentence aggregation being an object, into groups whose semantic contents are similar. CONSTITUTION:With regard to a simple sentence group which has been selected by a processing part 111 of a word level similarity calculating part 11 of a similarity calculating part 1 of a simple sentence classifying system, and has been brought to each frame expression, each corresponding similarity is derived, based on sinonymity and similar knowledge related to a verb stored in a verb knowledge storage part 13, and sinonymity and similar knowledge related to a noun stored in a noun knowledge storage part 14. Its result is stored in a storage part 112, and a processing part 121 of a sentence level similarity calculating 12 takes notice of its stored result, fetches the similarity related to each role of a word group, and derives the similarity between frames. By this calculating part 12, all simple sentence groups of a sentence aggregation are processed, and outputted to a similar group generating part 2. Subsequently, the generating part 2 generates an aggregation of the similar group constituted of simple sentences whose semantic contents are similar to each other.