Fuzzy ART研究及在银行信用风险评估中的应用

探讨了Fuzzy ART神经网络的聚类功能及其参数对网络的影响。提出了一种基于该聚类理论的银行信用风险评估聚类模型。采用ASP.NET+MS SQL Server 2000的B/S构架实现了银行信用风险评估系统。通过上市公司财务数据验证了聚类结果的有效性和合理性。     

Grouping parts based on geometrical shapes and manufacturing attributes using a neural network

This paper proposes a new part family classification system (IPFACS: Image Processing and Fuzzy ART based Clustering System), which incorporates image processing techniques and a modified Fuzzy ART neural network algorithm. IPFACS can classify parts based on geometrical shape and manufacturing attributes, simultaneously. With a proper reduction and normalization of an image data through the image processing methods and scaling capability in the modified Fuzzy ART, different types of geometrical shape data and manufacturing attribute data can be simultaneously classified in the same system. IPFACS has been tested for an example set of hypothetical parts. The results show that IPFACS provides a good feasible approach to form part families based on both geometrical shape and manufacturing attributes.     

Neural network-based design of cellular manufacturing systems

A neural network based on a competitive learning rule, when trained with the part machine incidence matrix of a large number of parts, classifies the parts and machines into part families and machine cells, respectively. This classification compares well with the classical clustering techniques. The steady state values of the activations and interconnecting strengths enable easier identification of the part families, machine cells, overlapping parts and bottleneck machines. Neural networks are mostly applied by treating them as a blackbox, i.e. the interaction with the environment and the information acquisition and retrieval occurs at the input and the output level of the network. This paper presents an approach where knowledge is extracted from the external and internal structure of the neural network.     

Constructive feedforward ART clustering networks. I

Part I of this paper proposes a definition of the adaptive resonance theory (ART) class of constructive unsupervised on-line learning clustering networks. Class ART generalizes several well-known clustering models, e.g., ART 1, improved ART 1, adaptive Hamming net (AHN), and Fuzzy ART, which are optimized in terms of memory storage and/or computation time. Next, the symmetric Fuzzy ART (S-Fuzzy ART) network is presented as a possible improvement over Fuzzy ART. As a generalization of S-Fuzzy ART, the simplified adaptive resonance theory (SART) group of ART algorithms is defined. Gaussian ART (GART), which is found in the literature, is presented as one more instance of class SART. In Part II of this work, a novel SART network, called fully self-organizing SART (FOSART), is proposed and compared with Fuzzy ART, S-Fuzzy ART, GART and other well-known clustering algorithms. Results of our comparison may easily extend to the ARTMAP supervised learning framework.     

加热炉钢坯温度软测量模型研究

研究基于模糊聚类的钢坯温度神经网络软测量模型.该方法由两个部分组成, FCM(Fuzzy C-Means)聚类算法用来对训练样本进行分类,分布式RBF(Radial Basis Function) 网络对每类样本进行训练.在线测量时,采用自适应模糊聚类算法对新的工况数据进行 隶属度计算.文中将该算法应用于步进式加热炉钢坯温度的预报,仿真结果表明该算法的有 效性.     

GOFAM: a hybrid neural network classifier combining fuzzy ARTMAP and genetic algorithm

Fuzzy ARTMAP (FAM), which is a supervised model from the adaptive resonance theory (ART) neural network family, is one of the conspicuous neural network classifier. The generalization/performance of FAM is affected by two important factors which are network parameters and presentation order of training data. In this paper we introduce a genetic algorithm to find a better presentation order of training data for FAM. The proposed method which is the combination of genetic algorithm with Fuzzy ARTMAP is called Genetic Ordered Fuzzy ARTMAP (GOFAM). To illustrate the effectiveness of GOFAM, several standard datasets from UCI repository of machine learning databases are experimented. The results are analyzed and compared with those from FAM and Ordered FAM which is used to determine a fixed order of training pattern presentation to FAM. Experimental results demonstrate the performance of GOFAM is much better than performance of Fuzzy ARTMAP and Ordered Fuzzy ARTMAP. In term of network size, GOFAM performs significantly better than FAM and Ordered FAM.     

Manufacturing cell formation with production data using neural networks

Batch type production strategies need adoption of cellular manufacturing (CM) in order to improve operational effectiveness by reducing manufacturing lead time and costs related to inventory and material handling. CM necessitates that parts are to be grouped into part families based on their similarities in manufacturing and design attributes. Then, machines are allocated into machine cells to produce the identified part families so that productivity and flexibility of the system can be improved. Zero-one part-machine incidence matrix (PMIM) generated from route sheet information is commonly presented as input for clustering of parts and machines. An entry of ‘1’ in PMIM indicates that the part is visiting the machine and zero otherwise. The output is generated in the form of block diagonal structure where each block represents a machine cell having more than one machines and a part family. The major limitations of this approach lies in the fact that important production factors like operation time, sequence of operations, and lot size of the parts are not accounted for. In this paper, an attempt has been made to propose a clustering methodology based on adaptive resonance theory (ART) for addressing these issues. Initially, a methodology considering only the operation sequence of the parts has been proposed. Then, the methodology is suitably modified to deal with combination of operation sequence and operation time of the parts to address generalized cell formation (CF) problem. A new performance measure is proposed to quantify the performance of the proposed methodology. The performance of the proposed algorithm is tested with benchmark problems from open literature and the results are compared with the existing methods. The results clearly indicate that the proposed methodology outperforms the existing methods in most cases.     

A bacteria foraging algorithm based cell formation considering operation time

The cellular manufacturing system (CMS) is considered as an efficient production strategy for batch type production. The CMS relies on the principle of grouping machines into machine cells and grouping machine parts into part families based on pertinent similarity measures. The bacteria foraging algorithm (BFA) is a new in development computation technique extracted from the social foraging behavior of Escherichia coli (E. coli) bacteria. Ever since Kevin M. Passino invented the BFA, one of the main challenges has been employment of the algorithm to problem areas other than those for which the algorithm was proposed. This research work inquires the first applications of this emerging novel optimization algorithm to the cell formation (CF) problem. In addition, a newly developed BFA-based optimization algorithm for CF is discussed. In this paper, an attempt is made to solve the cell formation problem meanwhile taking into consideration number of voids in cells and a number of exceptional elements based on operational time of the parts required for processing in the machines. The BFA is suggested to create machine cells and part families. The performance of the proposed algorithm is compared with a number of algorithms that are most commonly used and reported in the corresponding scientific literature such as similarity coefficients methods (SCM), rank order clustering (ROC), ZODIAC, GRAFICS, MST, GATSP, GP, K-harmonic clustering (KHM), K-means clustering, C-link clustering, modified ART1, GA (genetic algorithm), evolutionary algorithm (EA), and simulated annealing (SA) using defined performance measures known as modified grouping efficiency and grouping efficacy. The results lie in favor of better performance of the proposed algorithm.     

A self-organizing neural network approach for the design of cellular manufacturing systems

The crux problem of group technology (GT) is the identification of part families requiring similar manufacturing processes and the rearrangement of machines to minimize the number of parts that visit more than one machine cell. This paper presents an improved method for part family formation, machine cell identification, bottleneck machine detection and the natural cluster generation using a self-organizing neural network. In addition, the generalization ability of the neural network makes it possible to assign the new parts to the existing machine cells without repeating the entire computational process. A computer program is developed to illustrate the effectiveness of this heuristic method by comparing it with the optimal technique for large-scale problems.     

Study on semiparametric Wilcoxon fuzzy neural networks

Fuzzy neural network (FNN) has long been recognized as an efficient and powerful learning machine for general machine learning problems. Recently, Wilcoxon fuzzy neural network (WFNN), which generalizes the rank-based Wilcoxon approach for linear parametric regression problems to nonparametric neural network, was proposed aiming at improving robustness against outliers. FNN and WFNN are nonparametric models in the sense that they put no restrictions, except possibly smoothness, on the functional form of the regression function. However, they may be difficult to interpret and, even worse, yield poor estimates with high computational cost when the number of predictor variables is large. To overcome this drawback, semiparametric models have been proposed in statistical regression theory. A semiparametric model keeps the easy interpretability of its parametric part and retains the flexibility of its nonparametric part. Based on this, semiparametric FNN and semiparametric WFNN will be proposed in this paper. The learning rules are based on the backfitting procedure frequently used in semiparametric regression. Simulation results show that the semiparametric models perform better than their nonparametric counterparts.     

基于多传感器信息融合的球团密度检测系统

提出了一种基于多传感器融合的球团密度在线测量新方法。该方法以模糊神经网络技术为基础,综合考虑制团过程各因素对球团密度的影响,建立起球团密度的模糊神经网络融合模型。研制了以TI公司TMS320F240DSP为核心的检测系统,并进行了实验,实验结果表明检测系统输出与实验室分析测量值基本一致。     

一种ART2神经网络的改进算法

传统的ART2神经网络由于预处理阶段的归一化，易将重要但幅值较小的分量作为噪声清除，造成在分类中丢失重要信息，同时还存在模式漂移的不足，分析产生这些不足的原因，并基于去单位化以及类内样本与类中心的距离不同而对类中心偏移产生不同影响的思想，对传统的ART2神经网络算法进行了改进。对一组渐变数据的测试表明，改进后的网络有效改善了模式漂移现象。同时，改进的ART2神经网络在核辐射场数据处理分类中有一定的实用价值。     

使用LSA降维的改进ART2神经网络文本聚类

针对文本数据高维度的特点和聚类的动态性要求,结合隐含语义分析（LSA）降维,提出一种改进的ART2神经网络文本聚类算法,通过LSA凸显文本和词条之间的语义关系,减少无用噪声,降低数据维度和计算复杂性;采用改进的折中学习方法,减少计算步骤,加快ART2神经网络计算速度,并利用最近邻动态重组方法提高ART2网络聚类的稳定性,减弱算法对样本输入顺序的依赖。实验表明,改进的文本聚类算法能有效地实现动态文本聚类。     

基于语句聚类识别的知识动态提取方法研究

根据自然语言的群集现象和对知识体系动态更新的要求,该文提出了一种基于语句聚类识别的知识动态提取方法。文中首先给出了知识动态提取方法的研究框架,该框架描述了由自然语言文卷到面向对象知识体系的转换过程。研究了语句矢量化的有关问题,给出了若干基本定义和一个判定定理,讨论了句元属性矢量的后置处理。提出了基于神经网络的语句聚类识别方法,采用前信度概念作为语句识别结果可信性的一种度量,利用Matlab编写了一个ART2神经网络仿真程序,给出了该神经网络对语句识别效果且作了相应分析。根据ART2网络对语句进行识别的结果,需将聚类后的各个语句进行知识形式的转换,为此提出了中间代码生成的宽度优先方法,并定义后信度作为对语句识别及语义模型构造可信性 的一个最终评价指标;进而针对合取规则句型,具体介绍了该方法的实现步骤。最后运用结构建模新方法生成结构化的派生关系,从而完成了由自然语言文卷到面向对象知识体系的知识提取过程。作者将一个机械CAD为背景的应用实例贯穿全文,演示了该实例的具体实现,证实了所提方法的有效性。     

一种改进的ART2网络学习算法

分析了现有ART2网络存在的问题,提出了一种改进的ART2算法。该算法首先利用样本数据自身来初始化权值,然后按照同一类中的数据点到其聚类中心的距离之和越小(即类内偏差越小),聚类效果越好的原则来设计特征表示场和类别表示场之间的权值修正公式,最后通过比较输入样本和聚类中心的模来有效地利用模式的幅度信息。分析证明了该算法不仅能有效解决模式漂移问题、充分利用幅度信息,而且能提高聚类速度。     

基于ART 的RBF 网络结构设计

针对径向基函数(RBF)网络隐层结构难以确定的问题,基于自适应共振理论(ART)网络良好的在线分类特性,提出一种RBF网络结构设计算法。该算法将ART网络的聚类特性用于RBF网络结构设计中,通过对输入向量与已存模式的相似度比较将输入向量进行分类,确定隐含层节点个数和初始参数,使网络具有精简的结构。对典型非线性函数逼近的仿真结果表明,所提出的结构具有快速的学习能力和良好的逼近能力。     

Constructive feedforward ART clustering networks. II

For pt.I see ibid., p.645-61 (2002). Part I of this paper defines the class of constructive unsupervised on-line learning simplified adaptive resonance theory (SART) clustering networks. Proposed instances of class SART are the symmetric fuzzy ART (S-Fuzzy ART) and the Gaussian ART (GART) network. In Part II of our work, a third network belonging to class SART, termed fully self-organizing SART (FOSART), is presented and discussed. FOSART is a constructive, soft-to-hard competitive, topology-preserving, minimum-distance-to-means clustering algorithm capable of: 1) generating processing units and lateral connections on an example-driven basis and 2) removing processing units and lateral connections on a minibatch basis. FOSART is compared with Fuzzy ART, S-Fuzzy ART, GART and other well-known clustering techniques (e.g., neural gas and self-organizing map) in several unsupervised learning tasks, such as vector quantization, perceptual grouping and 3-D surface reconstruction. These experiments prove that when compared with other unsupervised learning networks, FOSART provides an interesting balance between easy user interaction, performance accuracy, efficiency, robustness, and flexibility     

基于幅值分量的ART2神经网络的改进

摘要： ART2神经网络由于其算法结构中固有的归一化环节,丢失了幅度信息,其相似量度是一种模式相位信息的量度,存在“同相位不可分”的缺点。文章针对此不足,将样本的幅度作为样本特征分量的办法,对传统的ART2网络进行了改进。实验证明,改进后ART2网络在处理集群分布样本时,性能优于传统ART2网络,同时,改进的ART2网络在核辐射场数据处理分类中有一定的实用价值     

Fuzzy Clustering Using A Compensated Fuzzy Hopfield Network

Hopfield neural networks are well known for cluster analysis with an unsupervised learning scheme. This class of networks is a set of heuristic procedures that suffers from several problems such as not guaranteed convergence and output depending on the sequence of input data. In this paper, a Compensated Fuzzy Hopfield Neural Network (CFHNN) is proposed which integrates a Compensated Fuzzy C-Means (CFCM) model into the learning scheme and updating strategies of the Hopfield neural network. The CFCM, modified from Penalized Fuzzy C-Means algorithm (PFCM), is embedded into a Hopfield net to avoid the NP-hard problem and to speed up the convergence rate for the clustering procedure. The proposed network also avoids determining values for the weighting factors in the energy function. In addition, its training scheme enables the network to learn more rapidly and more effectively than FCM and PFCM. In experimental results, the CFHNN method shows promising results in comparison with FCM and PFCM methods.     

An improved neural network for manufacturing cell formation

With structures inspired by the structure of the human brain and nervous system, neural networks provide a unique computational architecture for addressing problems that are difficult or impossible to solve with traditional methods. In this paper, an unsupervised neural network model, based upon the interactive activation and competition (IAC) learning paradigm, is proposed as a good alternative decision-support tool to solve the cell-formation problem of cellular manufacturing. The proposed implementation is easy to use and can simultaneously form part families and machine cells, which is very difficult or impossible to achieve by conventional methods. Our computational experience shows that the procedure is fairly efficient and robust, and it can consistently produce good clustering results.