Research Of Information Retrieval Based On Rough Set Theory

Rough set theory is a kind of new data analysis method to deal with vagueness and uncertainty. A rapid growth of interest can be lately seen in rough set theory and its applications. Taking up with organizing and accessing information resources precisely, effectively and intelligently, information retrieval has been becoming current research hotspot of intelligent information processing. This thesis mainly aims to study on the problems of rough set theory in retrieval model and information processing technology.The thesis proposes a novel information retrieval model based on conditional probability relation and rough set, it also provides a solution to design classification method and represent relationship between classes. Defining conditional probability relation in index words space offers conveniences to mine and utilize similary relationship between index words, and gets ability of obtaining satisfactory retrieval results by adjusting the support degree.The study of rule acquisition applies distribution reduction in rules induction system, defines distribution core and distribution reduction of a rule, and proposes a kind of incremental learning method based on distribution reduction to acquire rule knowledge database. It deals with inconsistent samples, uses heuristic algorithm to gain concise rule knowledge which accords with decision table, and updates the existing rule knowledge database with incremental way when a new sample is added in decision information system. This method would be applicable to inconsistent decision tables, and the rule set would be consistent with the original information system.