Showing papers on "Rough set published in 1994"

"Rough Sets, Fuzzy Sets and Knowledge Discovery"

Abstract: An Overview of Knowledge Discovery in Databases: Recent Progress and Challenges.- Rough Sets and Knowledge Discovery: An Overview.- Search for Concepts and Dependencies in Databases.- Rough Sets and Concept Lattices.- Human-Computer Interfaces: DBLEARN and SystemX.- A Heuristic for Evaluating Databases for Knowledge Discovery with DBLEARN.- Knowledge Recognition, Rough Sets, and Formal Concept Lattices.- Quantifying Uncertainty of Knowledge Discovered from Databases.- Temporal Rules Discovery Using Datalogic/R+ with Stock Market Data.- A System Architecture for Database Mining Applications.- An Attribute-Oriented Rough Set Approach for Knowledge Discovery in Databases.- A Rough Set Model for Relational Databases.- Data Filtration: A Rough Set Approach.- Automated Discovery of Empirical Laws in a Science Laboratory.- Hard and Soft Sets.- Rough Set Analysis of Multi-Attribute Decision Problems.- Rough Set Semantics for Non-Classical Logics.- A Note on Categories of Information Systems.- On Rough Sets in Topological Boolean Algebras.- Approximation of Relations.- Variable Precision Rough Sets with Asymmetric Bounds.- Uncertain Reasoning with Interval-Set Algebra.- On a Logic of Information for Reasoning About Knowledge.- Rough Consequence and Rough Algebra.- Formal Description of Rough Sets.- Rough Sets: A Special Case of Interval Structures.- A Pure Logic-Algebraic Analysis of Rough Top and Rough Bottom Equalities.- A Novel Approach to the Minimal Cover Problem.- Algebraic Structures of Rough Sets.- Rough Concept Analysis.- Rough Approximate Operators: Axiomatic Rough Set Theory.- Finding Reducts in Composed Information Systems.- PRIMEROSE: Probabilistic Rule Induction Method Based on Rough Set Theory.- Comparison of Machine Learning and Knowledge Acquisition Methods of Rule Induction Based on Rough Sets.- AQ, Rough Sets, and Matroid Theory.- Rough Classifiers.- A General Two-Stage Approach to Inducing Rules from Examples.- An Incremental Learning Algorithm for Constructing Decision Rules.- Decision Trees for Decision Tables.- Fuzzy Reasoning and Rough Sets.- Fuzzy Representations in Rough Set Approximations.- Trusting an Information Agent.- Handling Various Types of Uncertainty in the Rough Set Approach.- Intelligent Image Filtering Using Rough Sets.- Multilayer Knowledge Base System for Speaker-Independent Recognition of Isolated Words.- Image Segmentation Based on the Indiscernibility Relation.- Accurate Edge Detection Using Rough Sets.- Rough Classification of Pneumonia Patients Using a Clinical Database.- Rough Sets Approach to Analysis of Data of Diagnostic Peritoneal Lavage Applied for Multiple Injuries Patients.- Neural Networks and Rough Sets - Comparison and Combination for Classification of Histological Pictures.- Towards a Parallel Rough Sets Computer.- Learning Conceptual Design Rules: A Rough Sets Approach.- Intelligent Control System Implementation to the Pipe Organ Instrument.- An Implementation of Decomposition Algorithm and its Application in Information Systems Analysis and Logic Synthesis.- ESEP: An Expert System for Environmental Protection.- Author Index.

Anomaly detection

Abstract: Computer are finite discrete machines, the set of real numbers is an infinite continuum. So real numbers in computers are approximation. Rough set theory is the underlying mathematics. A 'computer' version of Weistrass theorem states that every sequence, within the radius of error, repeats certain terms infinitely many times. In terms of applications, the theorem guarantees that the audit trail has repeating patterns. Examining further, based on fuzzy-rough set theory, hidden fuzzy relationships (rules) in audit data are uncovered. The information about the repeating data and fuzzy relationships reflect "unconscious patterns" of users' habits. They are some deeper "signatures" of computer users, which provide a foundation to detect abuses and misuses of computer systems, A "sliding window information system" is used to illustrate the detection of a 'simple' virus attack. The complexity problem is believed to be controllable via rough set representation of data.

Dynamic Reducts as a Tool for Extracting Laws from Decisions Tables

Abstract: We apply rough set methods and boolean reasoning for knowledge discovery from decision tables. It is not always possible to extract general laws from experimental data by computing first all reducts [12] of a decision table and next decision rules on the basis of these reducts. We investigate a problem how information about the reduct set changes in a random sampling process of a given decision table could be used to generate these laws. The reducts stable in the process of decision table sampling are called dynamic reducts. Dynamic reducts define the set of attributes called the dynamic core. This is the set of attributes included in all dynamic reducts. The set of decision rules can be computed from the dynamic core or from the best dynamic reducts. We report the results of experiments with different data sets, e.g. market data, medical data, textures and handwritten digits. The results are showing that dynamic reducts can help to extract laws from decision tables.

Rough Classification with Valued Closeness Relation

Abstract: The problem being addressed is that of classification support using decision rules learned from examples. The learning methodology is based on the rough set theory which is particularly well suited to deal with inconsistencies in the set of examples. The rules produced using the rough set theory are categorized into deterministic and non-deterministic depending whether a condition part of the rule is uniquely related with a decision part or not. The classification support is performed by matching a new case to one of decision rules. A possible result is that the new case does not match any of the known rules. Then, a set of rules “nearest” to the description of the new case is presented to the decision maker. In order to find “nearest” rules, a distance measure based on a valued closeness relation, accepting both nominal and ordinal attributes, is used. A medical example illustrates the classification support.

Useful Feature Subsets and Rough Set Reducts

Abstract: In supervised classiication learning, one attempts to induce a classiier that correctly predicts the label of novel instances. We demonstrate that by choosing a useful subset of features for the indis-cernibility relation, an induction algorithm based on simple decision table can have high prediction accuracy on artiicial and real-world datasets. We show that useful feature subsets are not necessarily maximal independent sets (relative reducts) with respect to the label, and that, in practical situations , using a subset of the relative core features may lead to superior performance.

Tolerance rough sets, Čech topologies, learning processes

Abstract: Basic similarity relations concerning distances, colors, phonology, and synonyms are considered in order to motivate tolerance rough sets which, in contrast with usual rough sets, lead not only to an interval of approximations but to an operator that can be iterated and generates a hierarchy of objects that are only possible in the considered set A; another hierarchy concerns the position, with respect to A, of objects that surely are not in A. Tolerance rough sets are refined and generalized in their turn by means of Cech topologies permitting to approach the learning processes viewed as interactions among innate and acquired factors. The restriction of finiteness of information systems is abandoned, but problems of competition between relevance and cost are left open

Rough Relation Algebras

Abstract: Rough relation algebras were introduced by S. Comer as a generalisation of algebras of Pawlak's rough sets and Tarski's relation algebras. In this paper, some algebraic and arithmetical properties of rough relation algebras are studied and the representable rough relation algebras are characterised.

DBROUGH: A Rough Set Based Knowledge Discovery System

Abstract: Knowledge discovery in databases, or data mining, is an important objective in the development of data- and knowledge-base systems. An attribute-oriented rough set method is developed for knowledge discovery in databases. The method integrates learning from example techniques with rough set theory. An attribute-oriented concept tree ascension technique is first applied in generalization, which substantially reduces the computational complexity of the database learning processes. Then the rough set techniques are applied to the generalized relation to derive different knowledge rules. Moreover, the approach can find all the maximal generalized rules in the data. Based on these principles, a prototype database learning system, DBROUGH, has been constructed. Our study shows that attribute-oriented induction combined with rough set techniques provide an efficient and effective mechanism for knowledge discovery in database systems.

Induction of medical expert system rules based on rough sets and resampling methods.

Abstract: Automated knowledge acquisition is an important research issue in improving the efficiency of medical expert systems. Rules for medical expert systems consists of two parts: one is a proposition part, which represent a if-then rule, and the other is probabilistic measures, which represents reliability of that rule. Therefore, acquisition of both knowledge is very important for application of machine learning methods to medical domains. Extending concepts of rough set theory to probabilistic domain, we introduce a new approach to knowledge acquisition, which induces probabilistic rules based on rough set theory (PRIMEROSE) and develop a program that extracts rules for an expert system from clinical database, using this method. The results show that the derived rules almost correspond to those of medical experts.

Discovery of decision rules in relational databases: a rough set approach

Abstract: We develop an attribute-oriented rough set approach for the discovery of decision rules in relational databases. Our approach combines machine learning techniques and rough set theory. We consider a learning procedure to consist of the two phases data generalization and data reduction. In the data generalization phase, utilizing knowledge about concept hierarchies and relevance of the data, an attribute-oriented induction is performed attribute by attribute. Some undesirable attributes of the discovery task are removed and the primitive data in the databases are generalized to the desirable level; this process greatly decreases the number of tuples which must be examined for the discovery task and substantially reduces the computational complexity of the database learning processes. Subsequently, in data reduction phase, rough set theory is applied to the generalized relation; the cause-effect relationships among the condition and decision attributes in the databases are analyzed and the non-essential or irrelevant attributes to the discovery task are eliminated without losing information of the original database system. This process further reduces the generalized relation. Thus very concise and more accurate decision rules for each class in the decision attribute with little or no redundancy information, can be extracted automatically from the reduced relation during the learning process. Our study shows that attribute-oriented induction combined with rough set theory provide an efficient and effective mechanism for discovering decision rules in database systems.

Anomaly detection: a soft computing approach

Abstract: Computer are finite discrete machines, the set of real numbers is an infinite continuum. So real numbers in computers are approximation. Rough set theory is the underlying mathematics. A “computer” version of Weistrass theorem states that every sequence, within the radius of error, repeats certain terms infinitely many times. In terms of applications, the theorem guarantees that the audit trail has repeating patterns. Examining further, based on fuzzy-rough set theory, hidden fuzzy relationships (rules) in audit data are uncovered. The information about the repeating data and fuzzy relationships reflect “unconscious patterns” of user's habits. They are some deeper “signatures” of computer users, which provide a foundation to detect abuses and misuses of computer systems. A “sliding window information system” is used to illustrate the detection of a “simple” virus attack. The complexity problem is believed to be controllable via rough set representation of data.

A first-order logic for reasoning under uncertainty using rough sets

Abstract: Reasoning with uncertain information is a problem of key importance when dealing with knowledge from real situations. Obtaining the precise numbers required by many uncertainty-handling formalisms can be a problem when building real systems. The theory of rough sets allows us to handle uncertainty without the need for precise numbers, and so has some advantages in such situations. The authors develop a set of symbolic truth values based upon rough sets which may be used to augment predicate logic, and provide methods for combining these truth values so that they may be propagated when augmented logic formulae are used in automated reasoning.

Use of Rough Sets and Spectral Data for Building Predictive Models of Reaction Rate Constants

Abstract: A model for predicting the log of the rate constants for alkaline hydrolysis of organic esters has been developed with the use of gas-phase mid-infrared library spectra and a rule-building software system based on the mathematical theory of rough sets. A diverse set of 41 esters was used as training compounds. The model is an advance in the development of a generalized system for predicting environmentally important reactivity parameters based on spectroscopic data. By comparison to a previously developed model using the same training set with multiple linear regression (MLR), the rough-sets model provided better predictive power, was more widely applicable, and required less spectral data manipulation. [For the previous MLR model, a standard error of prediction (SEP) of 0.59 was calculated for 88% of the training set data under leave-one-out cross-validation. In the present study using rough sets, an SEP of 0.52 was calculated for 95% of the data set.] More importantly, analysis of the decision rules generated by rough-sets analysis can lead to a better understanding of both the reaction process under study and important trends in the spectral data, as well as underlying relationships between the two.

Fuzzy sets and secure computer systems

Abstract: Computer are finite discrete machines, the set of real numbers is an infinite continuum. So real numbers in computers are approximation. Rough set theory is the underlying mathematics. A “computer” version of Weistrass theorem states that every sequence, within the radius of error, repeats certain terms infinitely many times. In terms of applications, the theorem guarantees that the audit trail has repeating patterns. Examining further, based on fuzzy-rough set theory, hidden fuzzy relationships (rules) in audit data are uncovered. The information about the repeating data and fuzzy relationships reflect “unconscious patterns” of user's habits. They are some deeper “signatures” of computer users, which provide a foundation to detect abuses and misuses of computer systems. A “sliding window information system” is used to illustrate the detection of a “simple” virus attack. The complexity problem is believed to be controllable via rough set representation of data.

Fuzzy set quantification of roughness in a rough relational database model

Abstract: This paper provides the theoretical background and important definitions for rough sets. It defines fuzzy rough sets and provides some basic properties of union and intersection of fuzzy rough sets. It then informally presents the basic properties of the rough relational database model. Because the fuzzy rough relational database (FRRD) is an extension of the basic rough relational database model, only a short, informal description of the basic model is provided. The FRRD model is described providing necessary definitions and discussion of operators. Advantages of the enhanced model are discussed. >

Rough Set Analysis of Multi-Attribute Decision Problems

Abstract: Rough set theory answers two basic questions related to multi-attribute decision problems: one about explanation of a decision situation and, another, about prescription of some decisions basing on analysis of a decision situation. In this paper, four classes of multi-attribute decision problems are defined, depending on a structure of their representation, their interpretation and the kind of questions related, as well as the rough set methodology of their analysis are briefly described.

Logics Systems for Approximate Reasoning: Approximation via Rough Sets and Topological Spaces

Abstract: Mathematicians formalized the approximation in terms of topology. In this paper a new family of logic systems for approximate reasoning, called Near Logic, is proposed; their semantics are rested on the notion of neighborhood system-a building block of topology. Somewhat surprisingly, the axiom schema of the Near Logic is that of the modal logic S4. This generalizes the fact that the axiom schema of Rough Logic is S5. The agreement in geometric and modalic considerations seems indicate that the proposed approach must have captured some intrinsic meaning of the approximate reasoning. Neighborhood Systems are very general approximation, so Near Logic can be regarded as a small step toward the formalization what Hao Wang called ”approximate proof” three decades ago.

Algebraic Specification of Empirical Inductive Learning Methods based on Rough Sets and Matroid Theory

Abstract: In order to acquire knowledge from databases, there have been proposed several methods of inductive learning, such as ID3 family and AQ family. These methods are applied to discover meaningful knowledge from large databases, and their usefulness is ensured. However, since there has been no formal approach proposed to treat these methods, efficiency of each method is only compared empirically. In this paper, we introduce matroid theory and rough sets to construct a common framework for empirical machine learning methods which induce the combination of attribute-value pairs from databases. Combination of the concepts of rough sets and matroid theory gives us an excellent framework and enables us to understand the differences and the similarities between these methods clearly. In this paper, we compare three classical methods, AQ, Pawlak's Consistent Rules and ID3. The results show that there exists the differences in algebraic structure between the former two and the latter and that this causes the differences between AQ and ID3.

Management of Uncertainty in AI: A Rough Set Approach.

Abstract: We present some consequences of the assumption that objects are classified on the basis of a partial information about them encoded in information systems. The presented results are based on the rough set approach [14] and boolean reasoning [1].

Fuzzy Set Theoretic Tools for Image Analysis

Abstract: Publisher Summary This chapter describes various fuzzy set theoretic tools and explores their effectiveness in representing/describing various uncertainties that might arise in an image-recognition system and the ways these can be managed in making a decision. In the chapter, some examples of uncertainties that arise often in the process of recognizing a pattern are discussed, and it describes various fuzzy set theoretic tools for measuring information on grayness ambiguity and spatial ambiguity in an image. The concepts of bound functions and spectral fuzzy sets for handling uncertainties in membership functions are also discussed in the chapter. Their applications to low-level vision operations whose outputs are crucial and responsible for the overall performance of a vision system are presented in the chapter for demonstrating the effectiveness of these tools in managing uncertainties by providing both soft and hard decisions. Their usefulness in providing the quantitative indices for autonomous operations is also explained in the chapter. The chapter also describes the issues of feature/primitive extraction, knowledge acquisition and syntactic classification, and the features of Dempster-Shafer theory and rough set theory in this context. An application of the multivalued recognition system for detecting curved structures from remotely sensed image is also described in the chapter.

Characterization of Relevance and Irrelevance in Empirical Learning Methods based on Rough Sets and Matroid Theory

Abstract: One of the most important characteristics of empirical learning methods, such as AQ, ID3(CART), C4.5 and CN2, is that they find variables which are relevant to classification. In this paper, we define relevance in empirical classifier as relevance of each given attribute to apparent or predictive classification, and describe this type of relevance in terms of rough sets and matroid theory. The results show that these algorithms can be viewed as the greedy algorithms searching for apparent classification and that their weight functions may play an important role in predictive classification.

Machine learning: rough sets perspective

Abstract: The paper presents a non-inductive, incremental technique for learning from examples derived within the context of the probabilistic variable Precision Rough Sets model. The technique involves the classification of the domain of interest into a relatively small number of categories followed by computation of all, or some, minimal rules with probabilities by using the concept of a decision matrix. >

Comparability and usefulness of newer and classical data analysis techniques. Application in medical domain classification

Abstract: In the present paper three different approaches to the classification of objects between several a priori known, distinct classes are compared. These are: rough sets, DISQUAL method as well as classification and regression trees (CART). A common set of medical data forms a basis for the comparison. AU the methods allow for a mixture of continuous and discrete attributes on input via discretization of continuous ones. The latter is done automatically in CART but should be given a priori in two others. Thus sensitivity on changes in discretization is also studied.