A new approach to attribute reduction of consistent and inconsistent covering decision systems with covering rough sets

Traditional rough set theory is mainly used to extract rules from and reduce attributes in databases in which attributes are characterized by partitions, while the covering rough set theory, a generalization of traditional rough set theory, does the same yet characterizes attributes by covers. In this paper, we propose a way to reduce the attributes of covering decision systems, which are databases characterized by covers. First, we define consistent and inconsistent covering decision systems and their attribute reductions. Then, we state the sufficient and the necessary conditions for reduction. Finally, we use a discernibility matrix to design algorithms that compute all the reducts of consistent and inconsistent covering decision systems. Numerical tests on four public data sets show that the proposed attribute reductions of covering decision systems accomplish better classification performance than those of traditional rough sets.     

An improved attribute reduction scheme with covering based rough sets

Attribute reduction is viewed as an important preprocessing step for pattern recognition and data mining. Most of researches are focused on attribute reduction by using rough sets. Recently, Tsang et al. discussed attribute reduction with covering rough sets in the paper (Tsang et al., 2008), where an approach based on discernibility matrix was presented to compute all attribute reducts. In this paper, we provide a new method for constructing simpler discernibility matrix with covering based rough sets, and improve some characterizations of attribute reduction provided by Tsang et al. It is proved that the improved discernibility matrix is equivalent to the old one, but the computational complexity of discernibility matrix is relatively reduced. Then we further study attribute reduction in decision tables based on a different strategy of identifying objects. Finally, the proposed reduction method is compared with some existing feature selection methods by numerical experiments and the experimental results show that the proposed reduction method is efficient and effective.     

The further investigation of covering-based rough sets: Uncertainty characterization, similarity measure and generalized models

The notion of rough sets was originally proposed by Pawlak. In Pawlak’s rough set theory, the equivalence relation or partition plays an important role. However, the equivalence relation or partition is restrictive for many applications because it can only deal with complete information systems. This limits the theory’s application to a certain extent. Therefore covering-based rough sets are derived by replacing the partitions of a universe with its coverings. This paper focuses on the further investigation of covering-based rough sets. Firstly, we discuss the uncertainty of covering in the covering approximation space, and show that it can be characterized by rough entropy and the granulation of covering. Secondly, since it is necessary to measure the similarity between covering rough sets in practical applications such as pattern recognition, image processing and fuzzy reasoning, we present an approach which measures these similarities using a triangular norm. We show that in a covering approximation space, a triangular norm can induce an inclusion degree, and that the similarity measure between covering rough sets can be given according to this triangular norm and inclusion degree. Thirdly, two generalized covering-based rough set models are proposed, and we employ practical examples to illustrate their applications. Finally, relationships between the proposed covering-based rough set models and the existing rough set models are also made.     

On axiomatic characterizations of three pairs of covering based approximation operators

Rough set theory is a useful tool for dealing with inexact, uncertain or vague knowledge in information systems. The classical rough set theory is based on equivalence relations and has been extended to covering based generalized rough set theory. This paper investigates three types of covering generalized rough sets within an axiomatic approach. Concepts and basic properties of each type of covering based approximation operators are first reviewed. Axiomatic systems of the covering based approximation operators are then established. The independence of axiom set for characterizing each type of covering based approximation operators is also examined. As a result, two open problems about axiomatic characterizations of covering based approximation operators proposed by Zhu and Wang in (IEEE Transactions on Knowledge and Data Engineering 19(8) (2007) 1131-1144, Proceedings of the Third IEEE International Conference on Intelligent Systems, 2006, pp. 444-449) are solved.     

复合覆盖粗糙集模型及其应用

针对覆盖粗糙集仅适用于单一数据类型的论域覆盖的问题,提出复合覆盖粗糙集模型。在研究邻域覆盖粗糙集、集值覆盖粗糙集、区间值覆盖粗糙集的基础上,在复合数据模型下,通过建立多种覆盖关系（邻域覆盖、集值覆盖、区间值覆盖等）,提出复合覆盖粗糙集模型,并给出复合覆盖粗糙集相关概念及性质。该模型适用于多种数据类型（符号数据、区间数据、集合数据、数值数据等）的论域覆盖问题,通过实例说明了该模型在复合信息系统中的应用,进一步加深对复合覆盖粗糙集相关概念的理解。     

Parameterized attribute reduction with Gaussian kernel based fuzzy rough sets

Fuzzy rough sets are considered as an effective tool to deal with uncertainty in data analysis, and fuzzy similarity relations are used in fuzzy rough sets to calculate similarity between objects. On the other hand in kernel tricks, a kernel maps data into a higher dimensional feature space where the resulting structure of the learning task is linearly separable, while the kernel is the inner product of this feature space and can also be viewed as a similarity function. It has been reported there is an overlap between family of kernels and collection of fuzzy similarity relations. This fact motivates the idea in this paper to use some kernels as fuzzy similarity relations and develop kernel based fuzzy rough sets. First, we consider Gaussian kernel and propose Gaussian kernel based fuzzy rough sets. Second we introduce parameterized attribute reduction with the derived model of fuzzy rough sets. Structures of attribute reduction are investigated and an algorithm with discernibility matrix to find all reducts is developed. Finally, a heuristic algorithm is designed to compute reducts with Gaussian kernel fuzzy rough sets. Several experiments are provided to demonstrate the effectiveness of the idea.     

Soft sets and soft rough sets

In this study, we establish an interesting connection between two mathematical approaches to vagueness: rough sets and soft sets. Soft set theory is utilized, for the first time, to generalize Pawlak’s rough set model. Based on the novel granulation structures called soft approximation spaces, soft rough approximations and soft rough sets are introduced. Basic properties of soft rough approximations are presented and supported by some illustrative examples. We also define new types of soft sets such as full soft sets, intersection complete soft sets and partition soft sets. The notion of soft rough equal relations is proposed and related properties are examined. We also show that Pawlak’s rough set model can be viewed as a special case of the soft rough sets, and these two notions will coincide provided that the underlying soft set in the soft approximation space is a partition soft set. Moreover, an example containing a comparative analysis between rough sets and soft rough sets is given.     

On the relation between rough set reducts and jumping emerging patterns

This paper presents the relations between rough set reducts and jumping emerging patterns. Observations are introduced formally and supported by brief examples. Furthermore, we propose practical applications of these observations to the minimal reduct problem and to testing whether a given attribute set is differentiating. We believe that our study can be expanded so as to include other types of reducts and emerging patterns.     

Discernibility matrix simplification for constructing attribute reducts

This paper proposes a reduct construction method based on discernibility matrix simplification. The method works in a similar way to the classical Gaussian elimination method for solving a system of linear equations. Elementary matrix simplification operations are introduced. Each operation transforms a matrix into a simpler form. By applying these operations a finite number of times, one can transform a discernibility matrix into one of its minimum (i.e., the simplest) forms. Elements of a minimum discernibility matrix are either the empty set or singleton subsets, in which the union derives a reduct. With respect to an ordering of attributes, which is either computed based on a certain measure of attributes or directly given by a user, two heuristic reduct construction algorithms are presented. One algorithm attempts to exclude unimportant attributes from a reduct, and the other attempts to include important attributes in a reduct.     

变精度模糊粗糙集的一种定义

模糊粗糙集模型同经典粗糙集模型类似,容易受到噪音数据的影响.针对该问题,受变精度粗糙集模型的启发,提出了变精度模糊粗糙集的概念.针对现有变精度模糊粗糙集模型尚不能满足一些基本性质的缺陷,重新定义了模糊近似空间中某一模糊集的β-下近似和β-上近似,该定义方式能够满足上述的基本性质.     

覆盖概率粗糙集的模糊性

在经典覆盖近似空间中定义了论域上任意元素x的最小子覆盖,基于任意元素的最小子覆盖给出了覆盖粗糙集上、下近似新的描述,进而给出了已有覆盖概率粗糙集模型在最小子覆盖意义下的描述。同时,以覆盖概率粗糙集的粗糙隶属函数为基础,应用经典模糊集熵的概念讨论了覆盖概率粗糙集模糊性的度量。     

Invertible approximation operators of generalized rough sets and fuzzy rough sets

This paper studies the classes of rough sets and fuzzy rough sets. We discuss the invertible lower and upper approximations and present the necessary and sufficient conditions for the lower approximation to coincide with the upper approximation in both rough sets and fuzzy rough sets. We also study the mathematical properties of a fuzzy rough set induced by a cyclic fuzzy relation.     

Supervised hybrid feature selection based on PSO and rough sets for medical diagnosis

Medical datasets are often classified by a large number of disease measurements and a relatively small number of patient records. All these measurements (features) are not important or irrelevant/noisy. These features may be especially harmful in the case of relatively small training sets, where this irrelevancy and redundancy is harder to evaluate. On the other hand, this extreme number of features carries the problem of memory usage in order to represent the dataset. Feature Selection (FS) is a solution that involves finding a subset of prominent features to improve predictive accuracy and to remove the redundant features. Thus, the learning model receives a concise structure without forfeiting the predictive accuracy built by using only the selected prominent features. Therefore, nowadays, FS is an essential part of knowledge discovery. In this study, new supervised feature selection methods based on hybridization of Particle Swarm Optimization (PSO), PSO based Relative Reduct (PSO-RR) and PSO based Quick Reduct (PSO-QR) are presented for the diseases diagnosis. The experimental result on several standard medical datasets proves the efficiency of the proposed technique as well as enhancements over the existing feature selection techniques.     

Generalized rough sets over fuzzy lattices

This paper studies generalized rough sets over fuzzy lattices through both the constructive and axiomatic approaches. From the viewpoint of the constructive approach, the basic properties of generalized rough sets over fuzzy lattices are obtained. The matrix representation of the lower and upper approximations is given. According to this matrix view, a simple algorithm is obtained for computing the lower and upper approximations. As for the axiomatic approach, a set of axioms is constructed to characterize the upper approximation of generalized rough sets over fuzzy lattices.     

On the structure of generalized rough sets

In this paper we consider some fundamental properties of generalized rough sets induced by binary relations on algebras and show that

1.

Any reflexive binary relation determines a topology.

2.

If θ is a reflexive and symmetric relation on a set X, then O={A⊆X|θ_-(A)=A} is a topology such that A is open if and only if it is closed.

3.

Conversely, for every topological space (X,O) satisfying the condition that A is open if and only if it is closed, there exists a reflexive and symmetric relation R such that O={A⊆X|R_-(A)=A}.

4.

Let θ be an equivalence relation on X. For any pseudo ω-closed subset A of X, θ₋(A) is an ω-closed set if and only if ω(x, x, … , x) ∈ θ₋(A) for any x ∈ X.

Moreover we consider properties of generalized rough sets.     

The algebraic structures of generalized rough set theory

Rough set theory is an important technique for knowledge discovery in databases, and its algebraic structure is part of the foundation of rough set theory. In this paper, we present the structures of the lower and upper approximations based on arbitrary binary relations. Some existing results concerning the interpretation of belief functions in rough set backgrounds are also extended. Based on the concepts of definable sets in rough set theory, two important Boolean subalgebras in the generalized rough sets are investigated. An algorithm to compute atoms for these two Boolean algebras is presented.     

A systematic study on attribute reduction with rough sets based on general binary relations

Attribute reduction is considered as an important preprocessing step for pattern recognition, machine learning, and data mining. This paper provides a systematic study on attribute reduction with rough sets based on general binary relations. We define a relation information system, a consistent relation decision system, and a relation decision system and their attribute reductions. Furthermore, we present a judgment theorem and a discernibility matrix associated with attribute reduction in each type of system; based on the discernibility matrix, we can compute all the reducts. Finally, the experimental results with UCI data sets show that the proposed reduction methods are an effective technique to deal with complex data sets.     

On fuzzy rough sets based on tolerance relations

Dubois and Prade (1990) [1] introduced the notion of fuzzy rough sets as a fuzzy generalization of rough sets, which was originally proposed by Pawlak (1982) [8]. Later, Radzikowska and Kerre introduced the so-called (I,T)-fuzzy rough sets, where I is an implication and T is a triangular norm. In the present paper, by using a pair of implications (I,J), we define the so-called (I,J)-fuzzy rough sets, which generalize the concept of fuzzy rough sets in the sense of Radzikowska and Kerre, and that of Mi and Zhang. Basic properties of (I,J)-fuzzy rough sets are investigated in detail.     

基于信息量的完备覆盖约简算法

覆盖粗糙集是Pawlak粗糙集的一种重要推广。类似于Pawlak粗糙集,约简也是覆盖粗糙集中的核心问题之一。通过引入覆盖族的信息量的概念,讨论了覆盖协调集、约简以及核的等价判定定理,同时对覆盖的重要性进行了度量;在此基础上,提出一种完备的启发式覆盖约简算法,它能够从搜索空间中逐步删除不重要覆盖,避免对其重要性的重复计算;最后,通过一个购房综合评价的实例说明了该算法的可行性与有效性。