INFORMATION RETRIEVAL SYSTEMS: ESTIMATION BY FUZZY LOGIC

Abstract

Natural language-based information retrieval systems demand specific estimation methods. In the article a number of estimation methods of these systems are presented, the structure of information retrieval system KASKAD is described and advantages of one of the possible variants of estimating search systems efficiency using the fuzzy logic method are given. Two criteria are used for quantitative evaluation of search system quality - the search completeness and search precision. The dependence of search system quality evaluation on the criterion of completeness is illustrated on the graph. The experiment showed that the information retrieval system evaluation is determined not only by its technical characteristics but by wording of the request for information.