Review of Information Retrieval

Abstract

The term ‘information retrieval’ has various shades of meaning. From a psychological viewpoint, information retrieval in the human brain is a routine task that one is doing at every moment. In a library, information retrieval is a process for finding, among an enormous number of books on the shelves, that single piece of information relevant to one’s purpose. In this monograph information retrieval refers, primarily, to methods and processes for searching relevant information out of bibliographic databases that contain extremely large numbers of documents. In this chapter we review methods and studies of information retrieval in the ordinary (nonfuzzy) sense, leaving consideration of fuzzy retrieval for later chapters. The topics covered in the present chapter are:

• 1. An example of a document database and a prototypical information retrieval system are introduced. Eventually the information retrieval system, which is of the ordinary type, will be generalized to encompass fuzzy sets.

• 2. Evaluation of the effectiveness of information retrieval has been a major subject of studies in bibliographic information. Here criteria for evaluation are briefly reviewed.

• 3. Studies on associative indices, such as thesauri and automatic generation of thesauri, are reviewed.

• 4. Current interest in bibliographic research is concentrated on bibliographic citations. We review citation studies: availability of citation indices in information retrieval, citation measures as indicators of scientific activities, and citation clustering techniques.

• 5. Weighted retrieval has been studied by several authors. The concept of weighted retrieval points us toward retrieval using fuzzy sets. A diagrammatic representation for weighted retrieval suggested by Heaps (1978) is reviewed. In chapter 7 we present a similar representation based on fuzzy set theory and compare the two diagrammatic methods.