Search

Interactive data mining focuses on efficient and effective human-computer interactions for data analysis purposes. An interactive system is an integration of a human user and a computer machine. ICS, an interactive classification system, is implemented to demonstrate the power of interactive data mining. The interaction is mutually beneficial to users and machines. This article describes the architecture of ICS, and introduces the main features of ICS in the entire data mining process. [ABSTRACT FROM AUTHOR]

Query expansion has been extensively studied as a technique for increasing information retrieval performance. However, due to the volume of documents available on the web, many of the techniques that have been successful in traditional information retrieval systems do not scale well to web information retrieval. We propose a new technique based on conceptual semantic theories, in contrast to the structuralist semantic theories upon which other techniques are based. The source of the query expansion information is the concept network knowledge base. Query terms are matched to those contained in the concept network, from which concepts are deduced and additional query terms are selected. In this paper, we describe the theoretical basis for this in-progress research, along with some preliminary results. [ABSTRACT FROM AUTHOR]

We discuss an ontological framework for approximation, i.e., to approximation of concepts and vague dependencies specified in a given ontology. The presented approach is based on different information granule calculi. We outline the rough-fuzzy approach for approximation of concepts and vague dependencies. [ABSTRACT FROM AUTHOR]

In this paper, we discuss the most suitable "representation granularity", keeping several types of discernibility including individually discernibility and class discernibility. In the traditional "reduction" sense, the goal is to find the smallest number of attributes such that they enable us to discern each tuple or each decision class. However, once we pay attention to the number of attribute values too, that is, the size of each attribute, another criterion is needed. Indeed, we should ask ourselves about which one is better in the following two situations: 1) we can discern them with a single attribute of size ten, and 2) we can do this with two attributes of size five. This study answers this question with some criteria. Especially, we deal with continuous attributes. If we evaluate this difference in the light of understandability, we may prefer the latter, because they give more simple descriptions. Such a combination of simple nominal description helps us as a language or as a Kansei representation. To do this, we propose some criteria and algorithms to find near-optimal solutions for those criteria. In addition, we show some results for some databases in UCI Machine Learning Repository. [ABSTRACT FROM AUTHOR]

In this paper, we introduce a discrete event control for Cat and Mouse example based on a paraconsistent logic program EVALPSN stable model computation. Predicting and avoiding control deadlock states are crucial problems in discrete event control systems. We show that the EVALPSN control can deal with prediction and avoidance of control dadlock states in the Cat and Mouse by defining general rules to represent the deadlock states in EVALPSN, and is much more flexible than the previous version of EVALPSN Cat and Mouse control. We also show how to translate the control properties of the Cat and Mouse into EVALPSN. Keywords: discrete event control, EVALPSN(Extended Vector Annotated Logic Program with Strong Negation), stable model, control deadlock, paraconsistent logic program. [ABSTRACT FROM AUTHOR]

In different fields data are presented under the form of Property Systems or Attribute Systems (i. e. Information Systems). In order to collect items linked together by attributes or properties we can use a number of techniques whose results range from exact classifications to different kinds of approximations. This range depends on the collecting operators and the characteristics of the Information System at hand. In this paper we discuss how to transform Information Systems in order to apply a well-funded set of operators and to improve their precision. [ABSTRACT FROM AUTHOR]

We examine methods of valued tolerance relations where the conventional methods based on rough sets are extended in order to handle incomplete information. The methods can deal with missing values probabilistically interpreted. We propose a correctness criterion to the extension of the conventional methods. And then we check whether or not the correctness criterion is satisfied in a method of valued tolerance relations. As a result, we conclude that the method does not satisfy the correctness criterion. Therefore, we show how to revise the method of valued tolerance relations so that the correctness criterion can be satisfied. [ABSTRACT FROM AUTHOR]

We discuss the problems of spatio-temporal reasoning in the context of hierarchical information maps and approximate reasoning networks (AR networks). Hierarchical information maps are used for representation of domain knowledge about objects, their parts, and their dynamical changes. They are constructed out of information maps connected by some spatial relations. Each map describes changes (e.g., in time) of states corresponding to some parts of complex objects. We discuss the details of defining relations between levels of hierarchical information maps as well as between parts satisfying some additional constraints, e.g. spatial ones. [ABSTRACT FROM AUTHOR]

This paper presents a novel method for finding interesting patterns from spatio-temporal data. First, we perform a pairwise comparison of spatio-temporal sequences using the multiscale matching, taking into account the requirements for multiscale observation. Next, we construct the clusters of sequences using rough-set based clustering technique. Experimental results on real soccer game records demonstrated that the method could discover some interesting pass patterns that may be associated with successful goals. [ABSTRACT FROM AUTHOR]

The decision theory is defined typically as the combination of utility theory and probability theory. In this paper we generalize the decision theory as the performance measure theory and uncertainty theory. Intelligent agents look for approximate optimal decisions under bounded resources and uncertainty. The $-calculus process algebra for problem solving applies the cost performance measures to converge to optimal solutions with minimal problem solving costs, and allows to incorporate probabilities, fuzzy sets and rough sets to deal with uncertainty and incompleteness. Keywords: resource-bounded reasoning, problem solving, decision theory, process algebra, uncertainty. [ABSTRACT FROM AUTHOR]

The relation of similarity is essential in understanding and developing frameworks for reasoning with vague and approximate concepts. There is a wide spectrum of choice as to what properties we associate with similarity and such choices determine the nature of vague and approximate concepts defined in terms of these relations. Additionally, robotic systems naturally have to deal with vague and approximate concepts due to the limitations in reasoning and sensor capabilities. Halpern [1] introduces the use of subjective and objective states in a modal logic formalizing vagueness and distinctions in transitivity when an agent reasons in the context of sensory and other limitations. He also relates these ideas to a solution to the Sorities and other paradoxes. In this paper, we generalize and apply the idea of similarity and tolerance spaces [2,3,4,5], a means of constructing approximate and vague concepts from such spaces and an explicit way to distinguish between an agent's objective and subjective states. We also show how some of the intuitions from Halpern can be used with similarity spaces to formalize the above-mentioned Sorities and other paradoxes. [ABSTRACT FROM AUTHOR]

Live Logic is an integrated approach for support of the learning and decision making in conditions of uncertainty. The approach covers both induction of probabilistic logical hypotheses from known examples and deduction of the plausible solution for an unknown case based on the inducted hypotheses. The induction method generalizes empirical data, discovering statistical patterns, expressed in logical language. The deduction method uses multidimensional ranking to reconcile contradictory patterns exhibited by a particular case. The method was applied on clinical data of the patients with prostate cancer who underwent prostatectomy. The goal was to predict biochemical failure based on the pre- and post- operative status of the patient. The patterns found by the method proved to be insightful from the pathologist's point of view. Most of them had been confirmed on the control dataset. In our experiments, the predictive accuracy of the Live LogicTM was also higher than that of other tested methods. [ABSTRACT FROM AUTHOR]

This paper discusses the degree of granularity and dependence of contingency tables from the viewpoint of linear algebra. From the results of determinantal divisors, it seems that the devisors provide information on the degree of dependencies between the matrix of the whole elements and its submatrices and the increase of the degree of granularity may lead to that of dependence. However, this paper shows that a constraint on the sample size of a contingency table is very strong, which leads to the evaluation formula where the increase of degree of granularity gives the decrease of dependency. [ABSTRACT FROM AUTHOR]

We consider two classification approaches. The metric-based approach induces the distance measure between objects and classifies new objects on the basis of their nearest neighbors in the training set. The rule-based approach extracts rules from the training set and uses them to classify new objects. In the paper we present a model that combines both approaches. In the combined model the notions of rule, rule minimality and rule consistency are generalized to metric-dependent form. An effective polynomial algorithm implementing the classification model based on minimal consistent rules has been proposed in [2]. We show that this algorithm preserves its properties in application to the metric-based rules. This allows us to combine this rule-based algorithm with the k nearest neighbor (k-nn) classification method. In the combined approach the rule-based algorithm takes the role of nearest neighbor voting model. The presented experiments with real data sets show that the combined classification model have the accuracy higher than single models. [ABSTRACT FROM AUTHOR]

We address the problem of determination of the size of the test set which can can guarantee statistically significant results in classifier error estimation and in selection of the best classifier from a given set. We focus on the case of the 0-1 valued loss function and we provide one and two sides optimal bounds for Validation (known also as Hold-Out Estimate and Train-and-Test Method). We also calculate the smallest sample size, necessary for obtaining the bound for given estimation accuracy and reliability of estimation, and we present the results in tables. Finally, we propose strategies for classifier design using the bounds derived. Keywords: Computational learning theory, Model Selection, Model Assessment, Hold-Out Estimate, Train-and-Test, Validation. [ABSTRACT FROM AUTHOR]

In this paper, Mind Evolutionary Algorithm (MEA) is introduced to correct the nonlinearity of voltage-controlled oscillator (VCO) in linear frequency modulation continuous wave (LFMCW) radar level gauge. Firstly, the frequency modulation (FM) voltage is divided into several subsections. By using fast Fourier transform (FFT) analysis for the beat frequency signals and distilling the characteristic of the spectrum, an evaluation function is constructed. Then MEA is applied to optimize the end coordinates of the subsections to achieve the nonlinear curve of FM voltage so as to compensate for the nonlinearity of VCO. Experiments show that the proposed method has good correction performance with no requirement of additional hardware and measuring equipment and is easy to apply. [ABSTRACT FROM AUTHOR]

The Estimation of Distribution Algorithm is a new class of population based search methods in that a probabilistic model of individuals are estimated based on the high quality individuals and used to generate the new individuals. In this paper we compute 1) some upper bounds on the number of iterations required for global convergence of EDA 2) the exact number of iterations needed for EDA to converge to global optima. [ABSTRACT FROM AUTHOR]

The work leading to this paper is semantic image classification. The aim is to evaluate contributions of clustering mechanisms to organize low-level features into semantically meaningful groups whose interpretation may relate to some description task pertaining to the image content. Cluster assignment reveals underlying structures in the data sets without requiring prior information. The semantic component indicates that some domain knowledge about the classification problem is available and can be used as part of the training procedures. Besides, data structural analysis can be applied to determine proximity and overlapping between classes, which leads to misclassification problems. This information is used to guide the algorithms towards a desired partition of the feature space and establish links between visual primitives and classes. It derives into partially supervised learning modes. Experimental studies are addressed to evaluate how unsupervised and partially supervised fuzzy clustering boost semantic-based classification capabilities. [ABSTRACT FROM AUTHOR]

A method of constructing a classifier that uses fuzzy reasoning is described in this paper. Rules for this classifier are obtained by means of algorithms relying on a tolerance rough sets model. Got rules are in so called sharp" form, a genetic algorithm is used for fuzzification of these rules. Presented results of experiments show that the proposed method allows getting a smaller rules set with similar (or better) classification abilities. [ABSTRACT FROM AUTHOR]

This paper is concerned with describing and analyzing the control actions which are accomplished by a human operator, who controls a complex dynamic system. The decision model is expressed by means of a decision table with fuzzy attributes. Decision tables are generated by the fuzzification of crisp data, basing on a set of fuzzy linguistic values of the attributes. A T-similarity relation is chosen for comparing the elements of the universe. Fuzzy partitions of the universe with respect to condition and decision attributes are generated. The task of stabilization of the aircraft's altitude performed by a pilot is considered as an illustrative example. The limit-based and mean-based variable precision fuzzy rough approximations are determined. The measure of u-approximation quality is used for evaluating the consistency of the human operator's decision model, and assessing the importance of particular condition attributes in the control process. [ABSTRACT FROM AUTHOR]

A generalization of the original idea of rough sets and variable precision rough sets is introduced. This generalization is based on the concept of absolute and relative rough membership. Similarly to variable precision rough set model, the generalization called parameterized rough set model, is aimed at modeling data relationships expressed in terms of frequency distribution rather than in terms of a full inclusion relation used in the classical rough set approach. However, differently from variable precision rough set model, one or more parameters modeling the degree to which the condition attribute values confirm the decision attribute value, are considered. The properties of this extended model are investigated and compared to the classical rough set model and the variable precision rough set model. [ABSTRACT FROM AUTHOR]

In this paper, we study variable precision rough set models based on multiple decision tables. The models can control the admissible level of classification error in each table, the ratio of supporting decision tables to all decision tables and the ratio of opposing decision tables to all decision tables. As the classical rough set model plays a key role in analysis of decision tables such as reduction, rule induction, etc., the proposed variable precision rough set models will play a key role in analysis of multiple decision tables. [ABSTRACT FROM AUTHOR]

The article introduces the basic ideas and investigates the probabilistic version of rough set theory. It relies on both classification knowledge and probabilistic knowledge in analysis of rules and attributes. One-way and two-way inter-set dependency measures are proposed and adopted to probabilistic rule evaluation. A probabilistic dependency measure for attributes is also proposed and demonstrated to have the monotonicity property. This property makes it possible for the measure to be used to optimize and evaluate attribute based-representation through computation of attribute reduct, core and significance factors. [ABSTRACT FROM AUTHOR]

This paper focuses on a rough set method to analyze human evaluation data with much ambiguity such as sensory and feeling data. In order to handle totally ambiguous and probabilistic human evaluation data, we propose a probabilistic approximation based on information gains of equivalent classes. Furthermore, we propose a two-stage method to simply extract uncertain if-then rules using decision functions of approximate regions. Finally, we applied the proposed method to practical human sensory evaluation data and examined the effectiveness of the proposed method. The result shown that our proposed rough set method is more applicable to human evaluation data. [ABSTRACT FROM AUTHOR]

The rough relational database model was developed for the management of uncertainty in relational databases. In this paper we discuss rough functional dependencies and the normalization process used with them. Normalization is an important part of the relational database design process and rough normalization provides similar benefits for the rough relational database model. Keywords: normalization, functional dependency, rough sets, relational database, uncertainty. [ABSTRACT FROM AUTHOR]

Minimal rule generation in Non-deterministicInformationSystems (NISs), which follows rough sets based rule generation in DeterministicInformationSystems (DISs), is presented. According to certainrules and possiblerules in NISs, minimalcertainrules and minimalpossiblerules are defined. Discernibilityfunctions are also introduced into NISs for generating minimal certain rules. Like minimal rule generation in DISs, the condition part of a minimal certain rule is given as a solution of an introduced discernibility function. As for generating minimal possible rules, there may be lots of discernibility functions to be solved. So, an algorithm based on an order of attributes is proposed. A tool, which generates minimal certain and minimal possible rules, has also been implemented. Keywords: Rough sets, Non-deterministic information, Minimal rules, Discernibility functions, Tool for rule generation. [ABSTRACT FROM AUTHOR]

Rough Set Theory is a mathematical tool to deal with vagueness and uncertainty. Rough Set Theory uses a single information table. Relational Learning is the learning from multiple relations or tables. This paper studies the use of Rough Set Theory and Variable Precision Rough Sets in a multi-table information system (MTIS). The notion of approximation regions in the MTIS is defined in terms of those of the individual tables. This is used in classifying an example in the MTIS, based on the elementary sets in the individual tables to which the example belongs. Results of classification experiments in predictive toxicology based on this approach are presented. Keywords: Rough Set Theory, Variable Precision Rough Sets, multi-table information system, relational learning, prediction. [ABSTRACT FROM AUTHOR]

In incomplete data missing attribute values may be universally interpreted in several ways. Four approaches to missing attribute values are discussed in this paper: lost values, "do not care" conditions, restricted "do not care" conditions, and attribute-concept values. Rough set ideas, such as attribute-value pair blocks, characteristic sets, characteristic relations and generalization of lower and upper approximations are used in these four approaches. A generalized rough set methodology, achieved in the process, may be used for other applications as well. Additionally, this generalized methodology is compared with other extensions of rough set concepts. [ABSTRACT FROM AUTHOR]

Data preprocessing is an essential step of the KDD process. It makes it possible to extract useful information from data. We propose two coefficients which respectively study the informational contribution of initial data in supervised learning and the intrinsic structure of initial data in not supervised one. These coefficients are based on Kappa coefficient. The confrontation of these two coefficients enables us to determine if feature construction is useful. We can present a system allowing the optimization of preprocessing step : feature selection is applied in all cases; then the two coefficients are calculated for the selected features. With the comparison of the two coefficients, we can decide the importance of feature construction. [ABSTRACT FROM AUTHOR]

A core in information system is a set of attributes globally necessary to distinct objects from different decision classes (i.e., the intersection of all reducts of the information system). A notion of a pairwise core (2-core), which naturally extends the definition of a core into the case of pairs of attributes is presented. Some useful features concerned with the graph representation of pairwise cores are discussed. The paper presents also practical application of the notion of 2-core. It is known that a core (if exists) may be used to improve the reduct finding methods, since there exist polynomial algorithms for core construction. The same may be proven for a 2-core, which may be also used for estimation of minimal reduct size. Keywords: reducts, core, pairwise core, reduct finding algorithms. [ABSTRACT FROM AUTHOR]

In the paper a greedy algorithm for construction of partial tests is considered. Bounds on minimal cardinality of partial reducts are obtained. Results of experiments with software implementation of the greedy algorithm are described. Keywords: partial cover, partial test, partial reduct, greedy algorithm. [ABSTRACT FROM AUTHOR]

In this paper, we introduce the notion of generalized algebraic lower (upper) approximation operator and give its characterization theorem. That is, for any atomic complete Boolean algebra with the set of atoms, a map is an algebraic lower approximation operator if and only if there exists a binary relation R on such that L = R-, where R- is the lower approximation defined by the binary relation R. This generalizes the results given by Yao. [ABSTRACT FROM AUTHOR]

The aim of this paper is to compare concept lattices and approximation spaces. For this purpose general approximation spaces are introduced. It is shown that formal contexts and information systems on one hand and general approximation spaces on the other could be mutually represented e.g. for every information system exists a general approximation space such that both structures determines the same indiscernibility relation. A close relationship between Pawlak's approximation spaces and general approximation spaces also holds: for each approximation space exists a general approximation space such that both spaces determine the same definable sets. It is shown on the basis of these relationships that an extent of the every formal concept is a definable set in some Pawlak's approximation space. The problem when concept lattices are isomorphic to algebras of definable sets in approximation spaces is also investigated. [ABSTRACT FROM AUTHOR]

The standard rough set theory has been introduced in 1982 [5]. In this paper we use a topological concepts to investigate a new definitions for the lower and upper approximation operators. This approach is a generalization for Pawlak approach and the generalizations in [2, 7, 10, 12, 13, 14, 15, 16]. Properties of the suggested concepts are obtained. Also comparison between our approach and previous approaches are given. In this case, we show that the generalized approximation space is a topological space for any reflexive relation. Keywords: Approximation operators, similarity relations, rough sets, topological space. [ABSTRACT FROM AUTHOR]

The rough approximations are considered in the context of multi-criteria classification problem where evaluations of objects on particular criteria and their assignments to decision classes are imprecise and given in the form of intervals of possible values. Within Dominance-based Rough Set Approach (DRSA), the lower and upper approximations reflect the inconsistencies with respect to dominance principle. In the considered case, also the interval assignments have to be taken into account. This requires a new formulation of the dominance principle. A possible solution to the problem consists in introducing the second-order rough approximations. The methodology based on these approximations preserves well-known properties of rough approximations, such as rough inclusion, complementarity, identity of boundaries and monotonicity. [ABSTRACT FROM AUTHOR]

We introduce a pair of rough set approximations in formal concept analysis. The proposed approximation operators are defined based on both lattice-theoretic and set-theoretic operators. The properties of the approximation operators are examined. Algorithms for attribute reduction and object reduction in concept lattices are presented. [ABSTRACT FROM AUTHOR]

We present a rough set approach to vague concept approximation within the adaptive learning framework. In particular, the role of extensions of approximation spaces in searching for concept approximation is emphasized. Boundary regions of approximated concepts within the adaptive learning framework are satisfying the higher order vagueness condition, i.e., the boundary regions of vague concepts are not crisp. There are important consequences of the presented framework for research on adaptive approximation of vague concepts and reasoning about approximated concepts. An illustrative example is included showing the application of Boolean reasoning in adaptive learning. Keywords: vagueness, rough sets, higher order vagueness, adaptive learning. [ABSTRACT FROM AUTHOR]

In this study, we discuss a concept of shadowed sets and elaborate on their applications. To establish some sound compromise between the qualitative Boolean (two-valued) description of data and quantitative membership grades, we introduce an interpretation framework of shadowed sets. Shadowed sets are discussed as three-valued constructs induced by fuzzy sets assuming three values (that could be interpreted as full membership, full exclusion, and uncertain). The algorithm of converting membership functions into this quantification is a result of a certain optimization problem guided by the principle of uncertainty localization. With the shadowed sets of clusters in place, discussed are various ideas of relational calculus on such constructs. We demonstrate how shadowed sets help in problems in data interpretation in fuzzy clustering by leading to the three-valued quantification of data structure that consists of core, shadowed, and uncertain structure. Keywords: shadowed sets, three-valued logic, fuzzy sets, principle of uncertainty localization, relational equations, data interpretation, fuzzy clustering, outliers, cores of clusters. [ABSTRACT FROM AUTHOR]