Tadeusz Radecki

Fuzzy set theoretical approach to document retrieval

Abstract The aim of a document retrieval system is to issue documents which contain the information needed by a given user of an information system. The process of retrieving documents in response to a given query is carried out by means of the search patterns of these documents and the query. It is thus clear that the quality of this process, i.e. the pertinence of the information system response to the information need of a given user depends on the degree of accuracy in which document and query contents are represented by their search patterns. It seems obvious that the weighting of descriptors entering document search patterns improves the quality of the document retrieval process. A mathematical apparatus which takes into consideration, in a natural manner, the fact that the grades of importance of the descriptors in document search patterns are of the continuum type, that is an apparatus adequate to the description of a retrieval system of documents indexed by weighted descriptors is—among known mathematical methods—the theory of fuzzy sets, formulated by L.A. Zadeh. It is the aim of this paper to present a new method of document retrieval based on the fundamental operations of the fuzzy set theory. We start by introducing basic notions, then the syntax and semantics of the proposed language for document retrieval will be given and an algorithm allocating documents to particular queries will be described and its properties discussed. The basic advantage of the use of the fuzzy set theory for document retrieval system description is that it takes into consideration, in a simple way, the differentiation of the importance of descriptors in document search patterns and the differentiation of the formal relevance grades of particular documents of an information system to a given query. Documents of the highest grades (in the given information system) of formal relevance to the given query may be retrieved by means of the application of simple operations of the fuzzy set theory.

Mathematical model of information retrieval system based on the concept of Fuzzy thesaurus

Abstract The object of this paper is to present a mathematical model of information retrieval system based on the concept of fuzzy thesaurus. For this purpose the mathematical methods of A. Zadehs theory of fuzzy sets have been used. The application of the theory of fuzzy sets to the description of information retrieval system makes it possible (1) to take into account the fact that the degrȩe of importance of the particular terms in the search patterns of the particular documents and information requrets is of a continuous character, (2) to formalize the relations between terms describing documents and information requests (fuzzy thesaurus). The use of simple operations of the theory of fuzzy sets makes it possible to retrieve documents with the highest degree of relevance to the information requests of the users of the information system.

Mathematical model of time-effective information retrieval system based on the theory of fuzzy sets

Abstract Search patterns of documents and information requests are their better or worse representatives only, so it is important to carry on examinations on possibilities of designing self-learning information retrieval systems. Another important question is to elaborate such an organization of document search pattern set as to obtain an acceptable response time of the information system to a given information request. A self-learning process of the proposed information system consists in the determination—on a set of document and information request search patterns—of the similarity relation according to L. A. Zadeh. The organization of a set of document search patterns proposed in the paper ensures the limitation of document search pattern set searching process—when retrieving a response to a given information request—to one (or several) subset from previously determined subsets. This makes the information system response time acceptable. The proposed information retrieval strategy is discussed in terms of fuzzy sets.

Generalized Boolean methods of information retrieval

Abstract In most operational information retrieval systems the Standard retrieval methods based on set theory and binary logic are used. These methods would be much more attractive if they could be extended to include the importance of various index terms in document representations and search request formulations, in addition to a weighting mechanism which could be applied to rank the retrieved documents. This observation has been widely recognized in the literature as such extended retrieval methods could provide the precision of a Boolean search and the advantages of a ranked output. However, a closer examination of all the reported work reveals that up to the present the only possible approach of sufficient consistency and rigorousness is that based on recently developed fuzzy set theory and fuzzy logic. As the concept of a fuzzy set is a generalization of the conventional notion of a set, the generalization of the information retrieval methods based on set theory and binary logic can be derived in a natural way. The present paper describes such generalized Boolean information retrieval methods. The presentation of each includes an outline of its advantages and disadvan-tages, and the relationships between each particular method and the corresponding Standard information retrieval method based on set theory and binary logic are also discussed. It has been shown that these Standard retrieval methods are particular cases of information retrieval methods based on the theory of fuzzy sets and fuzzy logic. The considerations concerning the information retrieval methods presented are illustrated by simple examples.

Similarity Measures for Boolean Search Request Formulations

The aim of this article is to propose a means of determining the similarity between search request formulations, which are Boolean combinations of descriptors. The similarity measures introduced can be used to cluster Boolean search request formulations. The search request formulation clusters can in turn be utilized in methods for clustering document representations. The initiation of the work on the methodology for determining similarity measures for Boolean search request formulations has been encouraged by hitherto performed experimental research into information retrieval systems with search request formulations in the form of sets of certain descriptors, i.e., in the same form as document representations. These experiments have shown a preponderance of those methods for clustering document representations that use previously formed clusters of search request formulations. We can expect that by using the proposed methodology for determining the similarity between Boolean combinations of descriptors, we might obtain similar results for the case of Boolean search request formulations. The introductory part of the article provides the background for undertaking the research reported. The underlying notions are then presented, and the proposed measures of similarity between Boolean search request formulations are justified and illustrated by examples. In addition, some experimental results concerning the suggested similarity measures are given.

Outline of a fuzzy logic approach to information retrieval

An information retrieval method based on fuzzy logic is presented. The method described takes into account in a straightforward way the varying importance of descriptors which reflect the content of the information system documents as well as the varying formal relevance grades of documents in relation to a given query. The use of the simple operations of fuzzy logic allows retrieval of documents with the highest grades of formal relevance (in a given information system).

New Approach to the Problem of Information System Effectiveness Evaluation.

Abstract The object of this paper is to present a new kind of approach to the problem of information system effectiveness evaluation as based on the theory of fuzzy sets. On the basis of this theory, the concepts of relevance and pertinence, which are the basic concepts used in determining the indices of information system effectiveness evaluation, have been defined. Assuming that in evaluating the effectiveness of information systems, one should consider separately the problem of quality evaluation of the transformation of the contents of documents and information requests into their search patterns and the problem of quality evaluation of the process of profile control of a document set of the information system, definitions have been given of parameters of quality evaluation of the transformation of the contents of documents and information requests into their search patterns with regard to a given information request as well as of parameters of quality evaluation of the process with regard to the whole set of information requests under examination. Besides, parameters of quality evaluation of the process of profile control of a document set of the information system have been defined. The parameters of effectiveness evaluation of information systems put forward in this paper take account of the fact that both evaluation of the relevance and evaluation of the pertinence of documents are of a continuous character.

A theoretical background for applying fuzzy set theory in information retrieval

The paper deals with information retrieval systems, and especially with document retrieval systems. Some of the postulates most desired to be met by possible approaches to document retrieval are given, and the extent to which they are fulfilled by the established retrieval methodologies is discussed. In particular, a considerable place is devoted to the probabilistic approach and the document retrieval methodologies based on vector space theory and cluster analysis. The examination of the widely known theoretical approaches to document retrieval has resulted in the conclusion that fuzzy set theory and fuzzy logic provide a basis for fulfilling most of those postulates in a direct manner and with sufficient consistency and rigorousness. Since the concept of a fuzzy set is a generalization of the conventional notion of a set, the generalization of the document retrieval methods based on set theory and binary logic can be derived in a natural way. At the end of the paper the main advantages of the document retrieval approach based on fuzzy set theory and fuzzy logic are specified and briefly discussed.

On the Inclusiveness of Systems for Retrieval Documents Indexed by Unweighted Descriptors.

Abstract One of the means of reducing the information retrieval time is by taking advantage of the property of inclusiveness of information retrieval systems. When one knows the system response to a query which is more general in relation to another query, then in an inclusive retrieval system in order to retrieve the response to the more specific query it suffices to limit the information retrieval process to the search of the system response to the more general query. This paper demonstrates the property of inclusiveness of document retrieval systems where documents are indexed by unweighted descriptors, and in which query search patterns are Boolean functions of descriptors (systems using the Inverted File Method, the Canonical Structure File Method, or the Sequential File Method). The paper presents three methods for determining a partial ordering relation on a set of Boolean search patterns of queries, implying a partial ordering on the set of the system responses to these queries and discusses the adequacy of each method depending on the information retrieval method used. The paper also proves the general theorems concerning the property of inclusiveness with regard to the class of information retrieval systems considered. Furthermore, the methods for calculating the degree of generality/specificity between queries are given. The utilization of the property of inclusiveness may considerably reduce the operating costs of information retrieval systems (particularly systems using the Sequential File Method, e.g. the SDI systems). Moreover, the results of the studies presented in the paper give the possibility of gradual narrowing or broadening in an automatic way of a given Boolean search pattern of a query, which is of vital importance for on-line information retrieval systems.