Framework Of Approximate Reasoning Research Articles

A Decision Tree-Initialised Neuro-fuzzy Approach for Clinical Decision Support

Apart from the need for superior accuracy, healthcare applications of intelligent systems also demand the deployment of interpretable machine learning models which allow clinicians to interrogate and validate extracted medical knowledge. Fuzzy rule-based models are generally considered interpretable that are able to reflect the associations between medical conditions and associated symptoms, through the use of linguistic if-then statements. Systems built on top of fuzzy sets are of particular appealing to medical applications since they enable the tolerance of vague and imprecise concepts that are often embedded in medical entities such as symptom description and test results. They facilitate an approximate reasoning framework which mimics human reasoning and supports the linguistic delivery of medical expertise often expressed in statements such as ‘weight low’ or ‘glucose level high’ while describing symptoms. This paper proposes an approach by performing data-driven learning of accurate and interpretable fuzzy rule bases for clinical decision support. The approach starts with the generation of a crisp rule base through a decision tree learning mechanism, capable of capturing simple rule structures. The crisp rule base is then transformed into a fuzzy rule base, which forms the input to the framework of adaptive network-based fuzzy inference system (ANFIS), thereby further optimising the parameters of both rule antecedents and consequents. Experimental studies on popular medical data benchmarks demonstrate that the proposed work is able to learn compact rule bases involving simple rule antecedents, with statistically better or comparable performance to those achieved by state-of-the-art fuzzy classifiers.

A granular computing framework for approximate reasoning in situation awareness

We present our results on the adoption of a set-theoretic framework for granular computing to situation awareness. The proposed framework guarantees a high degree of flexibility in the process of creation of granules and granular structures allowing to satisfy the wide variety of requirements for perception and comprehension of situations where some elements must be perceived per similarity, others per spatial proximity, some must be fused to improve their comprehension, and so on. A second value is the support for approximate reasoning in situation awareness. A granular structure in particular represents a snapshot of a situation, and is a building block for the development of tools and techniques to reason on situation in order to reduce situation awareness errors and accelerate the process of decision-making. To this purpose, we show a technique to support operators in the analysis of conformity between a recognized situation and an expected one. A third value is the fact that we can support operators in having rapid and indicative measures of how two situations, e.g. a recognized and a projected, may differ. A preliminary evaluation instantiating our approach with self-organizing maps is reported and discussed. The results are encouraging with respect to the capability of improving perception and comprehension of a situation, reducing comprehension errors and supporting projection of situations.

Tractable approximate deduction for OWL

Today's ontology applications require efficient and reliable description logic (DL) reasoning services. Expressive DLs usually have high worst case complexity while tractable DLs are restricted in terms of expressive power. This brings a new challenge: can users use expressive DLs to build their ontologies and still enjoy the efficient services as in tractable languages? Approximation has been considered as a solution to this challenge; however, traditional approximation approaches have limitations in terms of performance and usability. In this paper, we present a tractable approximate reasoning framework for OWL 2 that improves efficiency and guarantees soundness. Evaluation on ontologies from benchmarks and real-world use cases shows that our approach can do reasoning on complex ontologies efficiently with a high recall.

An integral with respect to probabilistic-valued decomposable measures

Several concepts of approximate reasoning in uncertainty processing are linked to the processing of distribution functions. In this paper we make use of probabilistic framework of approximate reasoning by proposing a Lebesgue-type approach to integration of non-negative real-valued functions with respect to probabilistic-valued decomposable (sub)measures. Basic properties of the corresponding probabilistic integral are investigated in detail. It is shown that certain properties, among them linearity and additivity, depend on the properties of the underlying triangle function providing (sub)additivity condition of the considered (sub)measure. It is demonstrated that the introduced integral brings a new tool in approximate reasoning and uncertainty processing with possible applications in several areas.

Participatory Learning of Propositional Knowledge

Our objective here is to extend the participatory learning paradigm (PLP) to environments in which we are interested in learning information and knowledge expressed in terms of declarative statements. We first recall the basic idea of participatory learning, which stresses the important role of what is already believed in all aspects of the learning process. We then discuss the representation of declarative-type binary knowledge within Zadeh's framework of approximate reasoning. We look at the approximate reasoning inference mechanism and its capability for weighted propositions. We introduce ideas such as consistency, compatibility, and commitment that are needed for our objective. We then provide a version of the PLP that is appropriate for the task of learning declarative knowledge. Central to this is the new updation algorithm that is introduced. We finally look at the dynamic performance of this framework. A particularly notable feature is the unlearning and then learning that takes place when the external environment changes.

Soundness Preserving Approximation for TBox Reasoning

Large scale ontology applications require efficient and robust description logic (DL) reasoning services. Expressive DLs usually have very high worst case complexity while tractable DLs are restricted in terms of expressive power. This brings a new challenge: can users use expressive DLs to build their ontologies and still enjoy the efficient services as in tractable languages. In this paper, we present a soundness preserving approximate reasoning framework for TBox reasoning in OWL2-DL. The ontologies are encoded into EL++ with additional data structures. A tractable algorithm is presented to classify such approximation by realizing more and more inference patterns. Preliminary evaluation shows that our approach can classify existing benchmarks in large scale efficiently with a high recall.

Theory of a-Truth Degrees in n-Valued Gödel Propositional Logic

In order to establish a graded reasoning mechanism and provide a possible framework for approximate reasoning in n-valued propositional logic, this paper introduces the concept of α-truth degrees of propositions in n-valued Gdel logical system by using the infinite product of uniformly distributed probability spaces of cardinal n. It is proved that the general inference rules with truth degrees hold, and a sufficient and necessary condition to judge α-tautology is obtained. Moreover, an intrinsic pseudo-metric on the set of propositions is defined by means of the similarity degree between propositions, which makes it possible to develop approximate reasoning in n-valued propositional logic. The graded method proposed in this paper lays a foundation for the algorithmic realization of approximate reasoning and serves as a guideline for the graded reasoning about knowledge.

Integrated semantics and logic metric spaces

The concept of tautology degree of an abstract proposition is proposed in this paper by means of integral of the corresponding function with respect to certain semantics over [0,1]. Then a pseudo-metric among abstract propositions is introduced and the R-logic metric space is proposed where R is a suitable implication operator. Finally, a new theoretical framework for approximate reasoning in the R-logic metric space is provided.

Dealing with software process deviations using fuzzy logic based monitoring

The work presented in this paper concerns the monitoring of software processes using a fuzzy logic based approach. The proposed monitoring for software processes focuses on the detection of deviations with respect to given models. The level of deviation is computed for different aspects of the process like progress, cost, structure (order between activities), etc. The proposed approach is implemented by the OMEGA environment (OMEGA stands for On-line Monitoring Environment: General and Adaptable). The environment provides a language for defining monitoring models. Such models are executed by the environment's engine. Fuzzy sets theory is used in the representation of the Information handled by monitoring systems, and possibility theory in the reasoning upon it. Fuzzy logic offers significant advantages over other approaches due to its ability to naturally represent uncertain and imprecise information. It equally provides a good framework for approximate reasoning. The use of the proposed monitoring environment is illustrated in the context of the Unified Software Development Process.

Measuring information in possibilistic logic

We provide an overview of the theory of approximate reasoning and discuss the measurement of information in this reasoning system using specificity. It is then shown how to represent the binary propositional logic in the framework of approximate reasoning. Using the measure of specificity we show how to measure the information contained in the propositions of binary logic. Our measure essentially measures the proportion of possibilities eliminated by the proposition. Next we turn to the possibilistic logic and show how to represent this within the framework of approximate reasoning. We again, using specificity, provide for a measure of information of propositions in this logic. Finally we turn to the issue of quantified statements and show how to represent general quantified statements involving predicates within these two logics. ©1999 John Wiley & Sons, Inc.

On consequence in approximate reasoning

ABSTRACT The purpose of this paper is to explore the notion of consequence in Approximate Reasoning. While logical consequence has an standard (semantic and syntactic) definition in Exact Logic, several different definitions of that concept can be considered in the Approximate Reasoning framework. In this paper some definitions are compared with consequences derived from inference methods used in Artificial Intelligence. We will show that some inference methods are not covered by that definitions. Finally a more general concept of Approximate Consequence is introduced covering a wide variety of such inference methods.

Hierarchical scheduling based on approximate reasoning — A comparison with ISIS

A hierarchical scheduling problem is considered. The hierarchy consists of five levels. Each level has an associated rule base. There are six basic input factors: (1) the priority of a job, (2) the earliness of the due date of a job, (3) the likelihood of the availability of resources for the work station, w/s(s), (4) the capability of w/s(s) to meet physical constraints, (5) the capability of w/s(s) to meet quality requirements, and (6) work in progress inventory costs. There are five consequents, one for each level: (1) suggested place of a job in the order of jobs, (2) the suggested place of w/s(s) in the order of w/s(s), (3) the (current) suggested place of a w/s(s) in the order of w/s(s), (4) the next current, suggested place of w/s(s) in the order of w/s(s), and (5) the suggested place of the queue arrangements. Both the input factors (antecedents) and the output factors (consequents) take on linguistic terms to express the qualitative assessments of a scheduling expert within the context of the specific factor. An approximate reasoning framework, based on interval valued fuzzy sets and a search tree heuristic with similarity measures, is executed for the analysis of the hierarchical scheduling problem using a fuzzy tool box coded in LISP that interacts with KEE for the representation of the knowledge base in a frame structure. The software model is tested with real life data used in M. Fox's ISIS model. A comparison of the two models based on tardiness measure shows that there are slight advantages to fuzzy set representation and approximate reasoning with respect to tardiness measure above and beyond the ‘user friendly’ nature of knowledge representation available with fuzzy sets.

A fuzzy expert system for a service centre of spare parts

Abstract This paper presents the design of an approximate reasoning framework for an expert system prototype for a service centre of spare parts, to which customers bring failed items for repair. The design development is fundamentally based upon an analysis of a queuing model associated with the service centre system problem. This queuing model provides a prerequisite insight and the knowledge about such a service system. The building process for the framework is described in a case study utilizing the queuing model, namely, M/M/c repair systems with spares. The objective here is to aid management in determining certain decision policies and the capacities which are critical to them. Within the approximate reasoning framework, the identification and the construction of the basic rules that contain uncertain (vague, ambiguous, fuzzy) linguistic terms are described, as well as the specification of the membership functions that represent the meaning of such linguistic terms. Consistency of rules is studied in accordance with the internal relationships between system variables. Approximate Analogical Reasoning with a tree search is used as the inference engine of the expert system. Approximate reasoning results are compared with analytical results.

THE FUNDAMENTAL THEOREM OF ULTRAPRODUCT IN PAVELKA'S LOGIC

AbstractIn [This Zeitschrift 25 (1979), 45‐52, 119‐134, 447‐464], Pavelka systematically discussed propositional calculi with values in enriched residuated lattices and developed a general framework for approximate reasoning. In the first part of this paper we introduce the concept of generalized quantifiers into Pavelka's logic and establish the fundamental theorem of ultraproduct in first order Pavelka's logic with generalized quantifiers. In the second part of this paper we show that the fundamental theorem of ultraproduct in first order Pavelka's logic is preserved under some direct product of lattices of truth values.

An expert system prototype for inventory capacity planning: An approximate reasoning approach

An approximate reasoning framework is suggested for the development of an expert system prototype to aid management in planning inventory capacities. The development is considered to be a stage that comes after the analysis of a stochastic model. Such a model would provide the requisite insight and knowledge about the inventory system under specific assumptions. As a consequence, the model builder(s) would act as expert(s). The restructuring process from the stochastic model into the approximate reasoning framework is described in a case study analysis for a Markovian production model. The stochastic model considers a relatively simplified production process: one machine, constant production rate, a compound Poisson demand process for the product together with the reliability feature comprising the machine failure process and the ensuing repair action. In this context, the authors propose an approximate reasoning framework and describe (1) the identification of the managerial decision-making rules, which usually contain uncertain (vague, ambiguous, fuzzy) linguistic terms; and (2) the specification of membership functions that represent the meaning of such linguistic terms within context-dependent domains of concern. They then define a new universal logic incorporating these rules and functions and apply it to inventory capacity planning. Two case examples and a simulation experiment consisting of 21 cases are summarized with a discussion of results.

Concept recognition: An approximate reasoning framework

This article presents a knowledge based methodology for recognizing concept instances in complex data such as natural scenes or speech signals. the architecture of a prototype system performing this task is also described. In order to obtain the capability of handling noisy patterns, the knowledge representation is based on continuous valued logics, and the inference engine is able to do sophisticated reasoning about the knowledge it uses in order to take into account the possible degradation of cues in the pattern. The knowledge base is subdivided into a bulk knowledge and a degradation theory. the bulk knowledge consists of production rules capturing discriminating cues in the signal as they appear in the normal cases. the degradation theory describes how the cues can be degraded owing to insertion and deletion errors. The whole classification process consists of two phases. First of all, the rules of the bulk knowledge are applied to the pattern, trying to obtain a suitable classification, and the evidence assigned to the rules is combined with a technique based on Dempster-Shafer theory. In the second phase, the classification is refined taking into account also the possible degradations in the pattern. the resulting system is then able to deal with different kinds of uncertainty and errors, such as fluctuation in the values measured for the pattern and insertion-deletion errors. the method is illustrated and evaluated on a simple example of spoken word recognition. (This work has been developed within the pilot project ESPRIT No. 26. © 1989 Wiley Periodicals, Inc.)