Development of a Fuzzy Information Retrieval System

本课题旨在开发一套实用的模糊信息检索系统。为此，研究了模糊检索方法、快速检索算法和高级检索功能。具有模糊索引和模糊主题词的文献数据库已被开发出来。此外。模糊检索方法已应用于图片和DNA数据库的相似性检索。有关发展结果的更多详情载于下文。(1)方法发展模糊检索方法，如使用模糊词库的模糊关联检索，根据每个用户的偏好对文档使用优先权重的模糊过滤检索。提出了基于模糊命题指标的相似度检索方法。这些检索方法的快速算法已经被开发出来。(2)系统和数据库的开发开发了一个包含上述所有检索功能的模糊信息检索系统。这个系统可以模糊检索超过3万份文件记录。一个模糊同义词典已经开发了一个实际使用的大规模数据库。本文还开发了一个带有模糊索引和模糊词库的文档数据库。(3)应用模糊检索技术已被应用于图像数据库的相似性检索。建立了基于模糊命题指标的实验图像数据库，并实现了基于模糊集模型的相似度检索方法。提出了一种基于模糊集模型的匹配函数的DNA数据库相似性检索方法，并进行了检索实验。

The aim of the present research project is to develop a fuzzy information retrieval system for practical use. For this purpose, methodology of fuzzy retrieval, fast algorithms, and advanced functions of retrieval have been studied. Document databases with fuzzy indices and fuzzy thesaurs have been developed. Furthermore. fuzzy retrieval methods have been applied to similarity retrieval of pictures and a DNA database. More details of the results of the development are given in the following.(1) Methodological developmentMethods of fuzzy retrieval such as fuzzy association retrieval using fuzzy thesaurus, fuzzy filter retrieval that uses a prior weighting to documents according to each user's preference. and similarity retrieval using fuzzy propositional indices have been developed. Fast algorithms have been developed to each of these retrieval methods.(2) System and database developmentA fuzzy information retrieval system that includes all of the above retrieval functions has been developed. This system enables fuzzy retrieval of over 30, 000 document records. A fuzzy thesaurus has been developed for a large-scale database of practical use. A document database with fuzzy indices and a fuzzy thesaurus has also been developed.(3) ApplicationsFuzzy retrieval techniques have been applied to similarity retrieval of image databases. An experiental image database with a fuzzy propositional indices has been developed and similarity retrieval method using a fuzzy set model has been implemented. A similarity retrieval using a matching function based on a fuzzy set model has been developed for a DNA database and retrieval experiments have been performed.