A Method for Data Classification Based on Discernibility Matrix and Discernibility Function

A method for data classification will influence the efficiency of classification. Attributes reduction based on discernibility matrix and discernibility function in rough sets can use in data classification, so we put forward a method for data classification. Namely, firstly, we use discernibility matrix and discernibility function to delete superfluous attributes in formation system and get a necessary attribute set. Secondly, we delete superfluous attribute values and get decision rules. Finally, we classify data by means of decision rules. The experiments show that data classification using this method is simpler in the structure, and can improve the efficiency of classification.

An improved reduction algorithm based on the degree of attribute discernibility

This paper deals with the problem of attribute discernibility reduction and proposes some new concepts to rough set theory （RST） based on the discernibility matrix of Skowron, such as secondary core, regeneration matrix and the degree of attribute discernibility （DAD）. This paper puts forward an attribute reduction algorithm based on maximum discernibility degree, which opens up an effective way of gaining minimum attribute reduction of decision table. The efficacy of this algorithm has been verified by practical application in a diagnostic system of loader, which substantially decreases information gathering requirement and lowers the overall cost with no loss of accuracy.

Attribute reduction algorithm based on discernibility for decision table

The attribute reduction algorithms of decision table based on discernible matrix are required to construct discernible matrix, which reduces efficiency of algorithms. In this paper, the relationship between attribute discernible matrix and its discernibility is first established for general information systems. Based on the idea that the equivalent discernible matrix has a same attribute reduction, existing matrices are modified and a formula of attribute discernibility associated with algebraic reduction for decision table is proposed. A heuristic attribute reduction algorithm based on attribute discernibility is presented. Experimental results indicate that the algorithm can more easily explore an optimal or sub-optimal reduction, and is efficient.

Reduction Algorithms Based on Discernibility Matrix:The Ordered Attributes Method

In this paper, we present reduction algorithms based on the principle of Skowron's discernibility matrix - the ordered attributes method. The completeness of the algorithms for Pawlak reduct and the uniqueness for a given order of the attributes are proved. Since a discernibility matrix requires the size of the memory of U2, U is a universe of objects, it would be impossible to apply these algorithms directly to a massive object set. In order to solve the problem, a so-called quasi-discernibility matrix and two reduction algorithms are proposed. Although the proposed algorithms are incomplete for Pawlak reduct, their opimal paradigms ensure the completeness as long as they satisfy some conditions. Finally we consider the problem on the reduction of distributive object sets.

基于三支决策的差别矩阵属性约简算法

属性约简是粗糙集理论研究的核心内容之一,也是粗糙集理论的重要组成部分。该方法旨在减少冗余信息,提取出最具代表性和关键性质的属性集合。在属性约简的过程中,差别矩阵通常用于度量属性之间的关系,通过分析差别矩阵,研究者可以识别那些在描述系统行为方面贡献相似信息的属性,从而进行属性约简。基于三支决策的差别矩阵属性约简算法从差别矩阵的属性出发,首先刻画核以外的属性重要度,并以三支决策理论为基础构建一种新的属性约简方法。算法将传统概率粗糙集的上、下近似划分为三支决策中的正域、负域、边界域,基于不同的区域给出了决策规则,并通过决策损失函数来控制三支决策阈值。与同类算法相比,所提算法可以得到更为简洁的约简集和决策规则,且具有更小的时间复杂度。

基于依赖空间的对象导出三支概念格的属性约简

在给定形式背景下,基于依赖空间提出了对象导出三支概念格属性约简的新定义,在此基础上讨论了属性约简方法。首先,借助三支算子定义了对象幂集上的一致关系,引入依赖空间。通过依赖空间生成闭包算子的闭元素,研究了闭元素和对象导出三支概念的外延之间的关系。其次,提出基于依赖空间的对象导出三支概念格属性约简的定义,该定义下的约简集是保持由原属性集确定的一致关系不变的最小属性子集。最后,给出协调集的判定定理,利用可辨识属性矩阵及辨识函数给出对象导出三支概念格的属性约简方法。

Generalized binary discernibility matrix for attribute reduction in incomplete information systems

To extract and express the knowledge hidden in information systems, discernibility matrix and its extensions were introduced and applied successfully in many real life applications. Binary discernibility matrix, as a representative approach, has many interesting superior properties and has been rapidly developed to find intuitive and easy to understand knowledge. However, at present, the binary discernibility matrix is mainly adopted in the complete information system. It is a challenging topic how to achieve the attribute reduction by using binary discernibility matrix in incomplete information system. A form of generalized binary discernibility matrix is further developed for a number of representative extended rough set models that deal with incomplete information systems. Some useful properties and criteria are introduced for judging the attribute core and attribute relative reduction. Thereafter, a new algorithm is formulated which supports attribute core and attribute relative reduction based on the generalized binary discernibility matrix. This algorithm is not only suitable for consistent information systems but also inconsistent information systems. The feasibility of the proposed methods was demonstrated by worked examples and experimental analysis.

基于正域向量的决策粗糙集属性约简

在决策粗糙集模型中,现有划分层的正域都是通过集合求并所得,但基于该方法的保正域不变的属性约简与基于差别矩阵方法所得约简结果并不一致。提出了一种基于正域向量的决策粗糙集属性约简方法,该方法与基于差别矩阵的约简方法所得结果是一致的。最后给出一个算例说明其一致性。

Investigation on AQ11, ID3 and the Principle of Discernibility Matrix

The principle of discernibility matrix serves as a tool to discuss and analyze two algorithms of traditional inductive machine learning, AQ11 and ID3. The results are: (1) AQ11 and its family can be completely specified by the principle of discernibility matrix; (2) ID3 can be partly, but not naturally, specified by the principle of discernibility matrix; and (3) The principle of discernibility matrix is employed to analyze Cendrowska sample set, and it shows the weaknesses of knowledge representation style of decision tree in theory.

基于信息粒的不协调决策形式背景的分布约简

粒计算和知识约简是知识发现和数据挖掘的两个重要课题。基于信息粒研究不协调决策形式背景的属性约简的定义和方法。首先,利用对象概念的内涵定义对象集上的拟序关系,并研究其相关性质。然后,利用拟序类定义分布函数和最大部分函数,进而提出不协调决策形式背景的(最大)分布协调集和(最大)分布约简的定义。最后,定义不协调决策形式背景的(最大)分布辨识矩阵及(最大)分布辨识公式,基于辨识矩阵给出(最大)分布协调集的判定定理,并提出计算分布约简和最大分布约简的方法。

覆盖信息系统的一类悲观多粒度约简

在覆盖信息系统中,为探讨多粒度约简结构,文中利用辨识矩阵计算覆盖信息系统的悲观多粒度约简.首先,介绍覆盖信息系统中基于多粒度粗糙集的悲观属性约简定义.接着,构建辨识矩阵和辨识函数,通过对辨识矩阵进行析取和合取运算,可以精确地得到覆盖信息系统的所有悲观多粒度近似约简.最后,利用UCI的10个数据集进行了数值实验,验证了所提算法的可行性和有效性.

一个新的差别矩阵及其求核方法

首先利用反例指出HU的利用差别距阵来求粗糙集中的核的方法是错误的 ,然后给出一个新的差别距阵的定义和求核方法 ,并证明了方法的正确性 .

基于修正差别矩阵的高效属性约简算法

为降低基于修正差别矩阵的属性约简算法的复杂度,给出了基于修正差别矩阵的简化差别矩阵,证明了基于该简化差别矩阵的属性约简定义与基于原修正差别矩阵的属性约简定义是等价的.在此基础上设计了一个基于简化差别矩阵的属性约简算法,其空间和时间复杂度分别被降为O(|C|(|U′pos||U/C|))和max{O(|C|2(|Up′os||U/C|)),O(|C||U|log|U|)}.实例说明:用新算法进行属性约简,不仅减少了计算量,而且减少了存储空间,因而是一种高效的属性约简算法.

一个基于正区域的快速求核算法

现有利用差别矩阵设计的基于正区域的求核算法,其时间复杂度为O(|C‖U|2)。为降低求核算法的时间复杂度,给出了简化差别矩阵的定义和基于简化差别矩阵核的定义,并证明了该核与基于正区域的核是等价的。由于求简化差别矩阵的关键是求划分U/C,故利用基数排序的思想设计了一个快速求划分U/C的算法,其时间复杂度为O(|C‖U|)。在此基础上,利用简化差别矩阵设计了一个基于正区域的快速求核算法,其时间复杂度降为max{O(|C‖U|),O(|C‖U/C‖Up′os)}。实例说明了新算法的有效性。

基于差别矩阵的增量式属性约简完备算法

为了解决基于差别矩阵的属性约简完备算法得不到最小约简的问题,提出了一种改进的属性约简方法.该方法将信息论定义的属性重要性作为启发式信息,并通过构造一个条件信息熵算子对差别集合进行运算,同时利用算子来计算候选属性的剔除次序,采用宽度优先搜索策略使约简集合中含有最重要的属性,这样就解决了完备算法约简率低的问题.结合该方法并在分析对象集增量与差别矩阵关系的基础上,证明了增量约简定理,由此提出了一种增量式约简完备算法(CAIR),当新数据加入决策表时,算法可增量构造差别集合.实验结果表明,所提CAIR在大大缩短计算差别集合时间的同时,约简率比非完备算法提高了20.3%,是同条件下完备算法执行效率的13.2倍.

基于简化差别矩阵的增量式属性约简

新对象添加到决策表后,已有的属性约简将会发生改变,需要对其动态更新。为此,首先给出简化决策表和简化差别矩阵的定义,并证明了基于简化差别矩阵的属性约简与正区域的属性约简是等价的;然后,分析增量对象的不同情况,将增量属性约简映射到简化决策表上来实现,由此设计基于简化差别矩阵的增量式属性约简算法,并对算法进行改进;最后,利用实例和实验验证了所提出算法的正确性和高效性。

基于标记可辨识矩阵的增量式属性约简算法

针对现有增量式属性约简算法中存在的约简传承性差以及不完备现象,提出基于标记可辨识矩阵的增量式属性约简算法.本文首先定义了标记函数,对样本之间的可辨识性进行分类,并将之引入一个新的可辨识矩阵,在新增样本时,结合标记信息可以快速识别可辨识矩阵元素集的异动,获得强传承性的约简超集,在此基础上,设计与标记可辨识矩阵匹配的必要矩阵,用以快速判断并删除冗余属性,确保约简的完备性.理论分析以及实验测试表明,本算法具有约简传承性强,约简集完备等特点,具有较强的实用性.

一个基于差别矩阵的快速求核算法

给出简化差别矩阵和相应核的定义,并证明该核与差别矩阵的核是等价的。在此基础上设计了一个新的求核算法,使得新算法的时间复杂度和空间复杂度分别被降为max｛O(CU/C2),O(CU)｝和max｛O(U),O(CU/C2)｝。

一种新的求解属性值约简算法

属性约简和规则发现是数据挖掘研究的重要组成部分,本文提出了一种基于可分辨矩阵的属性值约 简新算法,并讨论了规则的提取方法,最后,通过对算法进行描述和实例验证论证了算法的有效性。

一种快速计算HU差别矩阵的属性约简算法

在已有的基于HU差别矩阵的属性约简算法中,一般是以差别矩阵中的元素作为启发信息而设计的,其时间复杂度为O(|C|2|U|2).为降低该属性约简算法的时间复杂度,首先引入简化决策表的定义,并设计了一个求简化决策表的算法,其时间复杂度为O(|C||U|).然后在简化决策表的基础上,定义了差别区域,并给出基于差别区域的属性约简定义,同时证明了基于差别区域的属性约简与基于差别矩阵的属性约简等价.在此基础上,以快速缩小简化决策表的搜索空间为目的,定义了一个新的、较为合理的、度量属性重要性的公式,并给出了它的递归计算方法,其时间复杂度为O(U/C|).最后以属性重要性为启发信息,设计了一个基于差别矩阵的快速属性约简算法,其时间复杂度降为max(O(|C||U|,O(|C|2|U/C|)),并用一个实例说明了新算法的高效性.理论分析与实验表明,新算法具有较好的扩展性.