PROGRESS OF THE RESEARCH OF INFORMATION-R ETRIEVAL MODELS IN THE 80S - A CRITICAL-REVIEW

In the research field of information retrieval(IR) models, there were many activities of research and development in the 80's. This paper re views articles primarily concerned with IR models, especially which ta ke quantitative approaches, for the purpose of tracing the progress in the research field critically. The number of those articles are too m any to review comprehensively in this paper, so the author chose a sub set of such articles written in Japanese or English. This paper, first of all, describes some basic problems of conventional IR systems base d on Boolean model. Secondly, it reviews several specific models in th e following order: vector space model, fuzzy set model, probabilistic model, integrated model, and other retrieval models. Finally, it concl udes that the progress of the field in the 80's can be summarized as c ontinuation and refinement of the results of the research activities i n the 60's and 70's, and that we cannot get any results which make a r adical change in the framework of the research field. However, this pa per also concludes that many useful, interesting results were produced in the same decade, and it suggests directions for further research w hich will bring fruitful results in the next decade.