Self-organizing Map Method Based on Real Rough Sets Space and Its Application of Pattern Recognition

This paper presents a real rough sets space and corresponding concepts of real lower and upper approximation sets which correspond to the real-valued attributes. Therefore, the real rough sets space can be investigated directly. A rhombus neighborhood for SOM is proposed, and the combination of SOM and rough sets theory is explored. According to the distance between the weight of winner node and the input vector in the real rough sets space, new weight learning rules are defined. The modified method makes the classification of the output of SOM clearer and the intervals of different classes larger. Finally, an example based on fault identification of an aircraft actuator is presented, The result of the simulation shows that this method is right and effective.

Minimal Critical Sets of Refined Inertias for Irreducible Sign Patterns of Order 2

Let S be a nonempty, proper subset of all refined inertias. Then, S is called a critical set of refined inertias for ireducible sign patterns of order n if is sufficient for any sign pattern A to be refined inertially arbitrary. If no proper subset of Sis a critical set of refined inertias, then S is a minimal critical set of refined inertias for sign patterns of order n . In this paper, all minimal critical sets of refined inertias for irreducible sign patterns of order 2 are identified. As a by-product, a new approach is presented to identify all minimal critical sets of inertias for irreducible sign patterns of order 2.

INERTIA SETS OF SYMMETRIC SIGN PATTERN MATRICES

A sign pattern matrix is a matrix whose entries are from the set {+,-,0}. The symmetric sign pattern matrices that require unique inertia have recently been characterized. The purpose of this paper is to more generally investigate the inertia sets of symmetric sign pattern matrices. In particular, nonnegative tri-diagonal sign patterns and the square sign pattern with all + entries are examined. An algorithm is given for generating nonnegative real symmetric Toeplitz matrices with zero diagonal of orders n≥3 which have exactly two negative eigenvalues. The inertia set of the square pattern with all + off-diagonal entries and zero diagonal entries is then analyzed. The types of inertias which can be in the inertia set of any sign pattern are also obtained in the paper. Specifically, certain compatibility and consecutiveness properties are established.

Inertia Set of a Nonnegative Symmetric Sign Pattern

For a symmetric sign pattern S1 the inertia set of S is defined to be the set of all ordered triples si（S） = {i（A） ： A = A^T ∈ Q（S）} Consider the n × n sign pattern Sn, where Sn is the pattern with zero entry （i,j） for 1 ≤ i = j ≤ n or｜i -j｜=n- 1 and positive entry otherwise. In this paper, it is proved that si（Sn） = {（n1, n2, n - n1 - n2）｜n1≥ 1 and n2 ≥ 2} for n ≥ 4.

基于Level Set方法的西夏字轮廓提取

随着国内外对西夏研究的不断深入,收藏于世界各地的大批西夏古籍文献通过影印方式陆续出版。如何将这些西夏古籍文献进行数字化、文本化则有着极其重要的意义。首先利用平滑和细化算法对西夏影印文献进行了预处理,然后利用Level set方法对影印文献中的西夏字进行了轮廓提取。Level Set演化函数在空间方向上采用了四阶紧致差分逼近式离散,计算过程中加入了窄带算法及全局优化方法。实验表明,算法在不增加计算时间的基础上可以得到较精确的西夏字轮廓。

A New Algorithm for Mining Frequent Pattern

Mining frequent pattern in transaction database, time series databases, and many other kinds of databases have been studied popularly in data mining research. Most of the previous studies adopt Apriori like candidate set generation and test approach. However, candidate set generation is very costly. Han J. proposed a novel algorithm FP growth that could generate frequent pattern without candidate set. Based on the analysis of the algorithm FP growth, this paper proposes a concept of equivalent FP tree and proposes an improved algorithm, denoted as FP growth * , which is much faster in speed, and easy to realize. FP growth * adopts a modified structure of FP tree and header table, and only generates a header table in each recursive operation and projects the tree to the original FP tree. The two algorithms get the same frequent pattern set in the same transaction database, but the performance study on computer shows that the speed of the improved algorithm, FP growth * , is at least two times as fast as that of FP growth.

Scaling up Kernel Grower Clustering Method for Large Data Sets via Core-sets

核栽培者是聚类最近 Camastra 和 Verri 建议的方法的一个新奇的核。它证明为各种各样的数据的好性能关于流行聚类的算法有利地设定并且比较。然而，方法的主要缺点是在处理大数据集合的弱可伸缩能力，它极大地限制它的应用程序。在这份报纸，我们用核心集合建议一个可伸缩起来的核栽培者方法，它是比为聚类的大数据的原来的方法显著地快的。同时，它能处理很大的数据集合。象合成数据集合一样的基准数据集合的数字实验显示出建议方法的效率。方法也被用于真实图象分割说明它的性能。

Fuzzy Condition Pattern Recognition and Fault Diagnosis for Spacecraft

This paper introduces concepts of symptom vector and fuzzy symptom vector forspacecraft condition recognition and fault diagnosis,defines an operator and suggests a fuzzy pat-tern recognition method of fault diagnosis for spacecraft.This method is verified by examples andresults are checked from an expert system.

A Fuzzy Neural Network for Fault Pattern Recognition

This paper combines fuzzy set theory with ART neural net-work , and demonstrates some important properties of the fuzzy ART neural net-work algorithm. The results from application on a ball bearing diagnosis indicate that a fuzzy ART neural net-work has an effect of fast stable recognition for fuzzy patterns.

Air Target Fuzzy Pattern Recognition Threat-Judgment Model

Threat-judgment is a complicated fuzzy inference problem. Up to now no relevant unified theory and measur-ing standard have been developed. It is very difficult to establish a threat-judgment model with high reliability in the airdefense system for the naval warships. Air target threat level judgment is an important component in naval warship com-bat command decision-making systems. According to the threat level judgment of air targets during the air defense of sin-gle naval warship, a fuzzy pattern recognition model for judging the threat from air targets is established. Then an algo-rithm for identifying the parameters in the model is presented. The model has an adaptive feature and can dynamicallyupdate its parameters according to the state change of the attacking targets and the environment. The method presentedhere can be used for the air defense system threat judgment in the naval warships.

Positional Information Storage in Sequence Patterns

We build a model of storage of well-defined positional information in probabilistic sequence patterns. Once a pattern is defined, it is possible to judge the effect of any mutation in it. We show that the frequency of beneficial mutations can be high in general and the same mutation can be either advantageous or deleterious depending on the pattern’s context. The model allows to treat positional information as a physical quantity, formulate its conservation law and to model its continuous evolution in a whole genome, with meaningful applications of basic physical principles such as optimal efficiency and channel capacity. A plausible example of optimal solution analytically describes phase transitions-like behavior. The model shows that, in principle, it is possible to store error-free information on sequences with arbitrary low conservation. The described theoretical framework allows one to approach from novel general perspectives such long-standing paradoxes as excessive junk DNA in large genomes or the corresponding G- and C-values paradoxes. We also expect it to have an effect on a number of fundamental concepts in population genetics including the neutral theory, cost-of-selection dilemma, error catastrophe and others.

基于词模式规则的轻量级日志模板提取方法

传统基于规则的日志解析方法针对每类日志需单独编写规则,且随着系统更新,出现新的日志模式时,需人工再次干预;基于深度学习的日志解析方法虽准确率高,但计算复杂度高。为解决日志解析方法人力成本和计算复杂度高的问题,文中提出一种基于词模式规则的轻量级日志模板提取方法,该方法由初始规则集生成、词模式规则应用、潜在错误样本发掘三个部分构成。首先,原始日志基于自适应随机抽样获取彼此间相似度较低的代表性日志;然后,基于专家反馈提取初始词模式规则集,在词模式规则应用模块对原始日志进行处理并提取日志模板;最后,在潜在错误样本发掘模块检查生成的日志模板聚类,发现潜在的错误分类样本并对其进行规则集更新。经过实验验证,在16个公开日志数据集上,文中方法的平均准确度达到97.8%,与基于深度学习的日志解析算法准确度基本持平;在计算效率方面,文中方法的单线程解析速度达到每秒20000条,且随着可用内核数量的增加,性能持续提升,满足系统日志的故障诊断和安全分析需求。

Rough similarity degree and rough close degree in rough fuzzy sets and the applications

Based on rough similarity degree of rough sets and close degree of fuzzy sets, the definitions of rough similarity degree and rough close degree of rough fuzzy sets are given, which can be used to measure the similar degree between two rough fuzzy sets. The properties and theorems are listed. Using the two new measures, the method of clustering in the rough fuzzy system can be obtained. After clustering, the new fuzzy sample can be recognized by the principle of maximal similarity degree.

基于Julia集的潮汕抽纱装饰纹样分形设计

分形普遍存在于自然环境的微观之中,在农业上常用于土壤特征与农作物结构等的分析。分形图案因其生于自然的特点,在解释自然结构的同时具有艺术性,与传统植物纹样搭配相适应。农作物作为潮汕抽纱纹样的主要题材,文章提出基于Julia集的潮汕抽纱植物装饰纹样分形设计模型,试图更好地实现潮汕抽纱数字化,提供符合当下审美的分形纹样设计,缓解自20世纪90年代工业化冲击后,抽纱从业人员凋零,缺乏从事其纹样设计人员的现状。通过对葡萄、麦穗、花卉等纹样进行归类与分形维度的计算,总结潮汕抽纱植物纹样的分形特征,应用分形艺术软件UF6完成分形纹样设计,与传统潮汕抽纱植物纹样遵循类别、形制进行组合。最后,针对青年群体进行模型输出结果的问卷调研,从民族性、创新性、艺术性、喜爱度验证分形引入潮汕抽纱的可行性,并邀请非遗工作者从生产角度做出评价。结果表明:潮汕抽纱可通过分形设计扩充新的纹样素材,青年群体希望潮汕抽纱能以局部装饰的形式融入日常应用的产品中。在生产上,根据应用于织品、印刷等不同需求,对分形纹样的复杂程度作出考量或简化,为刺绣织品图案的创新设计提供参考,展现分形通过农业所表现的艺术性。

基于特征约简与改进支持向量机的动态过程质量异常识别方法

为了有效降低特征维数并提高动态过程异常模式的识别精度,提出基于特征约简与改进支持向量机的动态过程质量异常识别方法。本文首先提取能反映质量异常模式的16个统计特征与7个形状特征,再使用粗糙集(RS)约简特征集合以剔除冗余特征与干扰特征。同时,使用遗传算法(GA)寻找支持向量机(SVM)的最优参数,并采用GA-SVM模型识别质量异常模式。仿真实验表明:粗糙集筛选后得到的12个特征具有较强区分动态过程是否出现异常状况的能力,遗传算法参数寻优后的支持向量机识别质量异常模式的精度明显高于其他同类型的模型,因此,本文提出的RS-GA-SVM模型具有良好的识别精度与稳健性,能够对动态过程进行有效监控。

新牧4号紫花苜蓿受精前后胚珠发育及种子形成动态

果实水平上的“胚珠多种子少”现象是种子生产中的一个共性问题,其中,有性繁殖过程至关重要,进行受精前后胚珠发育及种子形成动态的研究,对揭示种子产量的形成机制具有重要意义。以新牧4号紫花苜蓿为研究对象,通过切片、荧光观察及人工控制试验等,对其受精前后胚珠的发育、受精过程及结籽格局等进行研究,探讨有性繁殖方面制约种子产量形成的因素。结果表明:1)受精前,胚珠发育没见异常,腹缝线上也不具有维管分布的位置效应,但在受精过程中,19.6%的成熟胚囊被胼胝质包裹而不能正常受精。2)异花授粉较自花授粉具有更高的花粉萌发率和花粉管生长速度,进入子房后的花粉管优先达到子房的中部和基部,并使得这些位置的胚珠优先受精。3)自然条件下,授粉后第2天,中部和基部胚珠优先膨大,第3~6天,膨大胚珠的比例进一步增加,随后基部和中上部已经膨大的部分胚珠出现败育,最终形成荚果内两端胚珠败育率高、中间低的结籽格局。4)资源调控试验表明资源限制不仅影响受精前胚珠形成的数目,还加剧受精后已膨大胚珠的败育率。综上所述,新牧4号紫花苜蓿胚珠的发育与种子的形成受胼胝质沉积、受精顺序、资源分配状况及位置效应多种因素的影响,其中,由胼胝质沉积引起的受精率低以及资源限制引起的受精后胚珠的败育率高是影响种子产量的主要因素。

图片模糊集的一种相似度量及其在模式识别中的应用

图片模糊集能刻画具有模糊性、不确定性和不一致性的信息,相似度量是刻画两个对象的相似程度。文中研究图片模糊集的相似度量。考虑图片模糊集3个隶属度的信息差,基于指数函数构造出一种新的相似度量。文中提出的相似度量不仅满足相似度量的公理化定义,并且在实际应用中得到合理的计算结果。将相似度量应用到模式识别中,与现存的一些相似度量进行对比分析,结果表明所提出的相似度量不仅可以弥补一些现存相似度量的缺陷,还可以得到合理的计算结果。