基于k-最小表示误差类的表示分类方法

基于表示的分类(representation-based classification,RC)通常使用所有类的训练样本来表示测试样本.然而,是否需要使用全部类来表示测试样本仍有待研究.为此,提出一种两阶段表示分类框架.首先使用RC算法计算测试样本相对于全部类的训练样本的表示系数,找出前k(k≥1)个具有最小表示误差的类;然后利用该k个类的训练样本,再次应用RC算法对测试样本进行表示,并通过从这k个类中找出最小表示误差类来确定测试样本的类别.此外,提出了一种非负加权协同表示分类算法.所提分类框架中的前后两个RC算法可以相同也可以不同.取前后两个RC相同,对五种RC,在五个数据库上进行实验,实验结果表明,所提两阶段表示分类框架大多数情况下能显著提升原RC算法的分类精度.     

基于k-最小表示误差类的表示分类方法

基于表示的分类(representation-based classification,RC)通常使用所有类的训练样本来表示测试样本.然而,是否需要使用全部类来表示测试样本仍有待研究.为此,提出一种两阶段表示分类框架.首先使用RC算法计算测试样本相对于全部类的训练样本的表示系数,找出前k(k≥1)个具有最小表示误差的类;然后利用该k个类的训练样本,再次应用RC算法对测试样本进行表示,并通过从这k个类中找出最小表示误差类来确定测试样本的类别.此外,提出了一种非负加权协同表示分类算法.所提分类框架中的前后两个RC算法可以相同也可以不同.取前后两个RC相同,对五种RC,在五个数据库上进行实验,实验结果表明,所提两阶段表示分类框架大多数情况下能显著提升原RC算法的分类精度.     

一种新的拓展稀疏人脸识别算法

近几年来,基于稀疏表示分类是一个备受关注的研究热点。如果每类训练样本较充分,该类方法可以取得比较好的识别效果。当训练样本比较少时,它的分类效果可能就不理想。拓展的稀疏分类算法可以较好的解决这一问题,它在表示测试样本时,引入了训练样本的类内变量矩阵,利用它和训练样本集来表示测试样本,从而提高了人脸识别率。然而,该算法并没有考虑训练样本在表示测试样本中所起的作用,即所有训练样本的权重都等于1。本文采用高斯核距离对训练样本加权,提出用加权的训练样本和类内散度矩阵来共同表示测试样本,即基于加权的拓展识别算法。实验证明所提算法能够取得更好的人脸识别效果。     

基于鉴别性低秩表示的2阶段人脸识别算法

针对图像训练样本中存在噪声等情况,提出一种基于鉴别性低秩表示的2阶段人脸识别算法。该算法第1阶段是对所有训练样本进行低秩处理,筛选出M类与测试样本最相近的样本用于粗分类;第2阶段使用第1阶段筛选出来的样本做鉴别性低秩表示处理,并使用稀疏线性表示进行精细分类,决定测试样本最适合的类标签。本算法结合了低秩算法与稀疏算法的优点,在标准人脸库上的实验表明本算法表现优越。     

演化算法中一种基于单分类的预选择策略

演化算法中,预选择算子用于为后续的环境选择过程筛选出好的潜在候选后代解.现有预选择算子大多基于适应值评估、代理模型或分类模型.由于预选择过程本质上是一个分类过程,因此基于分类的预选择过程天然适用于演化算法.先前研究工作采用二分类或多分类模型进行预选择,需预先准备“好”和“差”两组或具有区分性的多组训练样本来构建分类模型,而随着演化算法的执行,“好”解和“差”解之间的界限将愈加模糊,因此准备具有区分性的两组或多组训练样本将变得具有挑战性.为解决该问题,本文提出了一种基于单分类的预选择策略(One-class Classification based PreSelection,OCPS),首先将当前种群中的解均视为“好”类样本,之后只利用该类“好”样本构建单分类模型,然后利用构建的模型对产生的多个候选解进行标记与选择.提出的策略应用在三个代表性演化算法中,数值实验结果表明,提出的策略能够提升现有演化算法的收敛速度.     

基于谱聚类欠取样的不均衡数据SVM分类算法

提出一种基于谱聚类欠取样的不均衡数据支持向量机(SVM)分类算法.该算法首先在核空间中对多数类样本进行谱聚类;然后在每个聚类中根据聚类大小和该聚类与少数类样本间的距离,选择具有代表意义的信息点;最终实现训练样本间的数目均衡.实验中将该算法同其他不均衡数据预处理方法相比较,结果表明该算法不仅能有效提高SVM算法对少数类的分类性能,而且总体分类性能及运行效率都有明显提高.     

一种基于核的模糊多球分类算法及其集成

提出了一种基于核的模糊多球分类算法,该算法在训练阶段为每一个模式类构造多个最小球覆盖其所有的训练样本,并且在识别阶段算法利用一个模糊隶属函数来归类测试样本。此外,在提出的分类算法的基础上,还给出了它的集成方法。最后,采用了4个真实数据集进行实验,实验结果表明该文提出的算法具有较好的分类性能,是一种行之有效的分类算法。     

核子类凸包样本选择方法及其SVM应用

提出一种基于核函数方法的类内训练样本选择方法——核子类凸包样本选择法,并将其用于支持向量机。该样本选择方法通过迭代方法,逐一选择了那些经映射后“距离已选样本”,并将其映射、生成“凸包最远的样本”。实验结果表明,该方法选择的少量样本使支持向量机获得了较高的识别比率,减少了存储需求,提高了分类速度。     

基于结构与文本关键词相关度的XML网页分类研究

针对XML网页特点，提出了计算XML文档结构相似性、文档关键词出现的位置以及关键词频度的方法，根据计算的结果提取XML网页特征，同时设计了一种基于支持向量机的XML网页多类分类算法．算法通过XML文档的训练样本集为每一类文档建立基于相似公共特征的聚类核，计算测试样本中的文档与每个聚类核的相似度，判断该文档的所属类．实验证明该分类算法具有比较高的分类查全率和查准率，能够较好地解决XML文档同时属于多个类的问题．     

基于聚类的加速k-近邻分类方法

实际生活中,经常会遇到大规模数据的分类问题,传统k-近邻k-NN(k-Nearest Neighbor)分类方法需要遍历整个训练样本集,因此分类效率较低,无法处理具有大规模训练集的分类任务。针对这个问题,提出一种基于聚类的加速k-NN分类方法 C_kNN(Speeding k-NN Classification Method Based on Clustering)。该方法首先对训练样本进行聚类,得到初始聚类结果,并计算每个类的聚类中心,选择与聚类中心相似度最高的训练样本构成新的训练样本集,然后针对每个测试样本,计算新训练样本集中与其相似度最高的k个样本,并选择该k个近邻样本中最多的类别标签作为该测试样本的预测模式类别。实验结果表明,C_k-NN分类方法在保持较高分类精度的同时大幅度提高模型的分类效率。     

“Booster method” for singularly perturbed robin problems i

An improved numerical method is presented for singularly perturbed Robin problems in this paper. In this method (“Booster method”), an asymptotic approximation is incorporated into a finite difference scheme to improve the numerical solution. Error estimates are proposed for the present method. In order to show the efficiency numerical examples are presented.     

Swarm intelligent approaches to auto-localization of nodes in static UWB networks

In this paper, we address the problem of localizing sensor nodes in a static network, given that the positions of a few of them (denoted as “beacons“) are a priori known. We refer to this problem as “auto-localization.” Three localization techniques are considered: the two-stage maximum-likelihood (TSML) method; the plane intersection (PI) method; and the particle swarm optimization (PSO) algorithm. While the first two techniques come from the communication-theoretic “world,” the last one comes from the soft computing “world.” The performance of the considered localization techniques is investigated, in a comparative way, taking into account (i) the number of beacons and (ii) the distances between beacons and nodes. Since our simulation results show that a PSO-based approach allows obtaining more accurate position estimates, in the second part of the paper we focus on this technique proposing a novel hybrid version of the PSO algorithm with improved performance. In particular, we investigate, for various population sizes, the number of iterations which are needed to achieve a given error tolerance. According to our simulation results, the hybrid PSO algorithm guarantees faster convergence at a reduced computational complexity, making it attractive for dynamic localization. In more general terms, our results show that the application of soft computing techniques to communication-theoretic problems leads to interesting research perspectives.     

Proximity Measures for the Classification of Geometric Figures

Abstract

A proximity relation is a fuzzy relation which is reflexive, symmetric, but not necessarily transitive. A quantitative measure of the proximity of two n-sided polygons is defined. Various properties of angular and dimensional proximities of triangles are investigated. A method for classifying a triangle as an “approximate right triangle,” “approximate isosceles triangle,” “approximate isosceles right triangle,” “approximate equilateral triangle” or “ordinary triangle” is presented. A method used to classify a quadrangle as “approximate square,” “approximate rectangle,” “approximate rhombus,” “approximate parallelogram,” “approximate trapezoid” or “ordinary quadrangle” is also presented. The measures of proximity employed for this purpose have an intuitive interpretation. The results may be of use in pattern recognition, information retrieval and artificial intelligence.     

Decentralized adaptive neural control of nonlinear interconnected large-scale systems with unknown time delays and input saturation

In this paper, a novel decentralized adaptive neural control scheme is proposed for a class of interconnected large-scale uncertain nonlinear time-delay systems with input saturation. RBF neural networks (NNs) are used to tackle unknown nonlinear functions, then the decentralized adaptive NN tracking controller is constructed by combining Lyapunov–Krasovskii functions and the dynamic surface control (DSC) technique along with the minimal-learning-parameters (MLP) algorithm. The stability analysis subject to the effect of input saturation constrains are conducted with the help of an auxiliary design system based on the Lyapunov–Krasovskii method. The proposed controller guarantees uniform ultimate boundedness (UUB) of all the signals in the closed-loop large-scale system, while the tracking errors converge to a small neighborhood of the origin. An advantage of the proposed control scheme lies in that the number of adaptive parameters for each subsystem is reduced to one, and three problems of “computational explosion”, “dimension curse” and “controller singularity” are solved, respectively. Finally, a numerical simulation is presented to demonstrate the effectiveness and performance of the proposed scheme.     

An application of fuzzy sets to the classification of geometric figures and chromosome images

The concept of a fuzzy set is applied to the classification of geometric figures and chromosome images through the use of shape-oriented angular and dimensional proximity measures. Various properties of approximate isosceles, approximate equilateral, approximate right, and approximate isosceles right triangles are investigated. A method for classifying a triangle as an “approximate right triangle,” “approximate isosceles triangle,” “approximate isosceles right triangle,” “approximate equilateral triangle,” or “ordinary triangle” is presented. A method used to classify a quadrangle as “approximate square,” “approximate rectangle,” “approximate rhombus,” “approximate parallelogram,” “approximate trapezoid,” or “ordinary quadrangle” is also presented. The measures of proximity employed for this purpose have an intuitive interpretation. The stochastic syntactic analysis technique and the “rubber-mask” technique have, in the past, been applied to the classification of chromosome images. In this paper, chromosome images are classified through the use of angular and dimensional proximity measures. The uttermost dimensional proximity and the least dimensional dissimilarity of chromosome images are defined and investigated. The results obtained in this paper may contribute to processing a picture from the polygonal approximation stage to the final classification stage in order to recognize a picture.     

“booster method” for singularly perturbed robin problems-II

Singularly perturbed Robin problems are considered in this paper. In order to get an improved numerical solution to these problems, a computational method (“Booster Method”) is suggested. In this method, an asymptotic approximation is incorporated into a finite difference scheme to improve the numerical solution. Error estimates are proposed when implemented in known difference schemes. Numerical examples are presented to illustrate the present method.     

基于“奖励制度”的DPoS共识机制改进

共识机制是区块链技术的核心。授权股权证明(Delegated Proof-of-Stake,DPoS)作为一种共识机制,其中每个节点都能够自主决定其信任的授权节点,从而实现快速共识验证。但DPoS机制仍然存在着节点投票不积极以及节点腐败的安全问题。针对这两个问题,文中提出了基于奖励的DPoS改进方案,投票奖励用以激励节点积极参与投票,举报奖励用以激励节点积极举报贿赂节点。Matlab仿真结果表明,投票奖励方法的引入提高了节点投票的积极性。与原始DPoS共识机制下投票节点数占比45%~50%相比,两种投票奖励方法使得参与投票节点数占比分别增加至65%~70%以及55%~60%。相比原始DPoS共识机制下不接受贿赂节点占比会随着恶意节点贿赂力度的加大而不断减少的情况,举报奖励方法的引入使得选择举报节点的占比出现了明显增加,在投票轮数为20的情况下,选择举报节点的总占比可以增至54%。实验结果表明,奖励制度的引入不但能够提高节点投票的积极性,而且增强了普通节点对恶意节点的贿赂抵抗性,使恶意节点成“代理人节点”的概率变小,保障了网络安全性。     

Adaptive Decentralized NN Control of Nonlinear Interconnected Time‐Delay Systems with Input Saturation

In this paper, a novel decentralized adaptive neural control scheme is proposed for a class of interconnected large‐scale uncertain nonlinear time‐delay systems with input saturation. Radial basis function (RBF) neural networks (NNs) are used to tackle unknown nonlinear functions. Then, the decentralized adaptive NN tracking controller is constructed by combining Lyapunov–Krasovskii functions and the dynamic surface control (DSC) technique, along with the minimal‐learning‐parameters (MLP) algorithm. The stability analysis subject to the effect of input saturation constraints are conducted with the help of an auxiliary design system based on the Lyapunov–Krasovskii method. The proposed controller guarantees uniform ultimate boundedness (UUB) of all of the signals in the closed‐loop large‐scale system, while the tracking errors converge to a small neighborhood around the origin. An advantage of the proposed control scheme lies in the number of adaptive parameters of the whole system being reduced to one and in the solution of the three problems of “computational explosion,” “dimension curse,” and “controller singularity”. Finally, simulation results along with comparisons are presented to demonstrate the advantages, effectiveness, and performance of the proposed scheme.     

An efficient and scalable pairwise key pre-distribution scheme for sensor networks using deployment knowledge

Pairwise key establishment is a fundamental security service for sensor networks. However, establishing pairwise key in sensor networks is a challenging problem, particularly due to the resource constraints on sensor nodes and the threat of node compromises. On the other hand, adding new nodes to a sensor network is a fundamental requirement for their continuous operation over time, too. We analyze the weaknesses of security due to node capture when adding sensor nodes using key pre-distribution schemes with “fixed” key pools. In this paper, we propose a new approach, which separates the nodes into groups, the nodes in a group communicate with each other using pairwise keys pre-distributed, the communications between any two neighbor groups are accomplished also through pairwise keys, which is computed based on the pre-distributed Hash chain. We show that the performance (e.g. continuous connectivity, continuous network resilience against node capture and memory usage) of sensor networks can be substantially improved by using our scheme. The scheme and its detailed performance evaluation are presented.