In this paper, a conceptual model based on fuzzy set techniques is proposed to handle the uncertainty in accessing information from petroleum exploration databases. One of the main contributions of/his paper is the development of a mathematical method to describe the retrieval process of petroleum exploration databases.


An information retrieval system is a software/hardware package that allows users to query and receive information that is stored in a datahase. If the database consists of document descriptors, a problem of particular interest is how best to match documents that arc relevant to a user's requests to items in the database. To meet this requirement it is necessary for the system to first understand precisely what the query means, and second to retrieve the documents that "best match" the query. An ideal information retrieval system should support users in formulating imprecisely specified queries, accepting them and handling intelligently. In order to deal with the uncertainty involved in the retrieval of information from petroleum exploration databases, a fuzzy matching method is proposed. A knowledge-based approach is used to incorporate some of the expert's tasks into the framework of the fuzzy retrieval process. This entails the construction of a knowledge base where semantic relations defined between object concepts in the domain arc explicitly represented. The knowledge base is conceptually partitioned into: l. the knowledge base storing object concepts and their relationships: 2. the knowledge base storing functions about numbers ranges of numbers and numbers related to time: 3. the knowledge base storing functions about qualitative concepts. Two binary relations - the synonym and implication relations between pairs of object concepts, arc used. Based on fuzzy set theory the suitable characteristic functions Ψrequired for handling different kinds of data types found in the petroleum exploration database arc proposed to determine the degree of support (DOS) for document d with respect to query q. We define a query as a triple Q = (Qc, Qn, Qq), where Qc deals with the domain-specific object concepts, Qn corresponds to numbers, ranges of numbers and the numbers related to time. Qq corresponds to the qualitative concepts. The concept of linguistic variables found in fuzzy logic is used to handle qualitative values. An algorithm, based on the Dempster-Shafer theory of evidence combination is proposed for computing a degree of similarity between object concepts in each document and the query. Finally, the retrieval process is illustrated by means of an example.

System architecture

We arc currently developing a retrieved system based on fuzzy model The system architecture shown in Figure I depicts a knowledge-based systems approach. The main functional units of the system are:

  • Natural Language Searching Interface

This interface incorporates fuzzy logic techniques to handle the uncertainty inherent in natural-language communication with information systems. This interface is capable of analyzing constrained natural-language queries and converting them to internal formats that are used by the inference mechanism to retrieve documents from the database. Specifically each query is viewed as a triple.

This content is only available via PDF.
You can access this article if you purchase or spend a download.