Evidential support logic programming

https://doi.org/10.1016/0165-0114(87)90110-2Get rights and content

Abstract

An expert system is a computer program which can act in a similar way to a human expert in a restricted domain of application from the point of view of solving problems, taking decisions, planning and giving advice. It consists of two parts. One part is a knowledge base consisting of that knowledge used by the expert in his performance. A second part is an inference engine which allows queries to be answered by asking questions of the environment and performing evidential reasoning.

This paper is concerned with the knowledge representation and inference mechanism for evidential reasoning. Man's knowledge consists of statements which cannot be guaranteed to be true and is expressed in a language containing imprecise terms. Uncertainties, either of a probabilistic or fuzzy nature, cannot be ignored when modelling human expertise. Not all practical reasoning takes the form of deductive inference. For practical affairs we use inductive, abductive, analogical and plausible reasoning methods and for each of these the concept of the strength of evidence would seem to be important.

We describe a support logic programming system which generalises logic programming to the case in which various forms of uncertainty can be included. In this system a conclusion does not logically follow from some axioms but is supported to a certain degree by means of evidence. The negation of the conclusion is also supported to a certain degree and the two supports do not necessarily add up to one.

A calculus for such a support logic programming system is described and applications to its use in expert systems and its use in providing recursive definitions of fuzzy concepts are given.

References (9)

There are more references available in the full text version of this article.

Cited by (120)

  • Proximity-based unification theory

    2015, Fuzzy Sets and Systems
    Citation Excerpt :

    This last approach is usually named Similarity-based Logic Programming, because the unification algorithm is supported by similarity relations. Examples of hybrid approaches, where both the resolution principle and the classical unification algorithm are modified, are [12] and [13,14]. In the case of [12] the fuzzy unification algorithm has a semantic nature, while in Baldwin's approach probabilistic and fuzzy uncertainty are combined into a single framework.

  • The operations on intuitionistic fuzzy values in the framework of Dempster-Shafer theory

    2012, Knowledge-Based Systems
    Citation Excerpt :

    For example, in [29] the author deals with a label space using mass assigned to the power set on the frame of discernment. The semantic of DST can be used in the frameworks of different approaches to the modelling uncertainty [6,7,26]. The above information of DST is quite enough for the presentation of A-IFS in terms of DST, but it is worth noting that besides the Definition 1 there are other possible representations of A-IFS proposed in the literature.

  • Identifying and eliminating dominated alternatives in multi-attribute decision making with intuitionistic fuzzy information

    2012, Applied Soft Computing Journal
    Citation Excerpt :

    Recently, more and more attention has been paid to multi-attribute decision making problems with intuitionistic fuzzy information. A lot of classical methods, such as the technique for order preference by similarity to ideal solution (TOPSIS) method [2], the maximizing deviation method [7], the elimination et choice translating reality (ELECTRE) method [8,9], the višekriterijumsko kompromisno rangiranje (VIKOR) method [10], and the Dempster–Shafer theory of evidence [11], have been extended to intuitionsitic fuzzy environments [12–20]. According to the information about the weight vector of attributes, the research on multi-attribute decision making can be mainly be summarized as: (1) intuitionistic fuzzy multi-attribute decision making with the completely known attribute weights [12,16,17,20–27]; (2) intuitionistic fuzzy multi-attribute decision making with the weight information on attributes completely unknown [13–15,28]; (3) intuitionistic fuzzy multi-attribute decision making with the partially known weight information on attributes [18,19,29–31]; and (4) dynamic intuitionistic fuzzy multi-attribute decision making [32].

  • Maximal confidence intervals of the interval-valued belief structure and applications

    2011, Information Sciences
    Citation Excerpt :

    The theory of belief functions, also called the Dempster–Shafer (DS) theory of evidence, was first introduced by Dempster [22] and was further mathematically formalized by Shafer [57] in the 1970s. In the 1980s, parallel formalizations continued with the work of Higashi and Klir [34], Hohle [35], Yager [87] and Baldwin [3]. In the following decade, Klir [39] further developed the DS theory and called it the “General Theory of Information”.

  • Reasoning with Uncertain and Conflicting Opinions in Open Reputation Systems

    2009, Electronic Notes in Theoretical Computer Science
View all citing articles on Scopus
View full text