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     

A survey of fuzzy clustering algorithms for pattern recognition. II

For pt.I see ibid., p.775-85. In part I an equivalence between the concepts of fuzzy clustering and soft competitive learning in clustering algorithms is proposed on the basis of the existing literature. Moreover, a set of functional attributes is selected for use as dictionary entries in the comparison of clustering algorithms. In this paper, five clustering algorithms taken from the literature are reviewed, assessed and compared on the basis of the selected properties of interest. These clustering models are (1) self-organizing map (SOM); (2) fuzzy learning vector quantization (FLVQ); (3) fuzzy adaptive resonance theory (fuzzy ART); (4) growing neural gas (GNG); (5) fully self-organizing simplified adaptive resonance theory (FOSART). Although our theoretical comparison is fairly simple, it yields observations that may appear parodoxical. First, only FLVQ, fuzzy ART, and FOSART exploit concepts derived from fuzzy set theory (e.g., relative and/or absolute fuzzy membership functions). Secondly, only SOM, FLVQ, GNG, and FOSART employ soft competitive learning mechanisms, which are affected by asymptotic misbehaviors in the case of FLVQ, i.e., only SOM, GNG, and FOSART are considered effective fuzzy clustering algorithms.     

Vector quantization of images using modified adaptive resonancealgorithm for hierarchical clustering

A modified adaptive resonance theory (ART2) learning algorithm, which we employ in this paper, belongs to the family of NN algorithms whose main goal is the discovery of input data clusters, without considering their actual size. This feature makes the modified ART2 algorithm very convenient for image compression tasks, particularly when dealing with images with large background areas containing few details. Moreover, due to the ability to produce hierarchical quantization (clustering), the modified ART2 algorithm is proved to significantly reduce the computation time required for coding, and therefore enhance the overall compression process. Examples of the results obtained are presented, suggesting the benefits of using this algorithm for the purpose of VQ, i.e., image compression, over the other NN learning algorithms.     

Models of self-organizing artificial neural networks to identify stationary industrial sources of air pollution

A problem of identifying one particular or a few possible pollution sources that are responsible for the deterioration of the air quality as a result of exceeding the standards of the maximum permissible emissions is considered. A model problem for a group of spatially divided stationary permanent industrial sources is solved. A statement identifying the problem and a method to solve it using two architectures of artificial neural networks, Kohonen’s networks for learning vector quantization with fixed and adaptive structures, as well as adaptive resonance theory network for analog inputs (ART-2), are presented. The method consists of clustering the data provided by self-learning algorithms (unsupervised learning). Estimation equations are given and operation algorithms of Kohonen’s and adaptive resonance theory networks at different life cycle stages are described. The results of the solution of the model problem that are obtained using each network is performed are comparatively analyzed.     

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.     

Incremental clustering of mixed data based on distance hierarchy

Clustering is an important function in data mining. Its typical application includes the analysis of consumer’s materials. Adaptive resonance theory network (ART) is very popular in the unsupervised neural network. Type I adaptive resonance theory network (ART1) deals with the binary numerical data, whereas type II adaptive resonance theory network (ART2) deals with the general numerical data. Several information systems collect the mixing type attitudes, which included numeric attributes and categorical attributes. However, ART1 and ART2 do not deal with mixed data. If the categorical data attributes are transferred to the binary data format, the binary data do not reflect the similar degree. It influences the clustering quality. Therefore, this paper proposes a modified adaptive resonance theory network (M-ART) and the conceptual hierarchy tree to solve similar degrees of mixed data. This paper utilizes artificial simulation materials and collects a piece of actual data about the family income to do experiments. The results show that the M-ART algorithm can process the mixed data and has a great effect on clustering.     

The modified fuzzy art and a two-stage clustering approach to cell design

This study presents a new pattern recognition neural network for clustering problems, and illustrates its use for machine cell design in group technology. The proposed algorithm involves modifications of the learning procedure and resonance test of the Fuzzy ART neural network. These modifications enable the neural network to process integer values rather than binary valued inputs or the values in the interval [0, 1], and improve the clustering performance of the neural network. A two-stage clustering approach is also developed in order to obtain an informative and intelligent decision for the problem of designing a machine cell. At the first stage, we identify the part families with very similar parts (i.e., high similarity exists in their processing requirements), and the resultant part families are input to the second stage, which forms the groups of machines. Experimental studies show that the proposed approach leads to better results in comparison with those produced by the Fuzzy ART and other similar neural network classifiers.     

Multi-objective optimization with an adaptive resonance theory-based estimation of distribution algorithm

The introduction of learning to the search mechanisms of optimization algorithms has been nominated as one of the viable approaches when dealing with complex optimization problems, in particular with multi-objective ones. One of the forms of carrying out this hybridization process is by using multi-objective optimization estimation of distribution algorithms (MOEDAs). However, it has been pointed out that current MOEDAs have an intrinsic shortcoming in their model-building algorithms that hamper their performance. In this work, we put forward the argument that error-based learning, the class of learning most commonly used in MOEDAs is responsible for current MOEDA underachievement. We present adaptive resonance theory (ART) as a suitable learning paradigm alternative and present a novel algorithm called multi-objective ART-based EDA (MARTEDA) that uses a Gaussian ART neural network for model-building and a hypervolume-based selector as described for the HypE algorithm. In order to assert the improvement obtained by combining two cutting-edge approaches to optimization an extensive set of experiments are carried out. These experiments also test the scalability of MARTEDA as the number of objective functions increases.     

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。     

ART2wNF及其稳定性——可塑性动态平衡性能分析

稳定性-可塑性两难问题的核心是系统如何在不削弱或忘记已学习模式的同时, 自适应地学习新事物. 目前公认自适应谐振理论(Adaptive resonance theory, ART) 能够部分解决稳定性-可塑性两难问题, 但依然存在学习受样本输入顺序影响大, 且存在学习中心渐变样本时, 带来的所谓模式漂移的问题. 受进化生物学关于人类学习的幼态延续特征的启发, 本文为每个F2层节点配备活跃度指示器λ, 并将其反馈回F1层参与STM(Short term memory) 向量的计算, 使这种新型ART2 网络在行为特征上具备幼态延续的显著特征, 本文称之为ART2wNF (Adaptive resonance theory with neoteny feature).论文从理论上证明算法的可行性, 并通过分析对随机生成样本集合的学习过程, 对比了ART2wNF 算法与常规ART2 网络在可塑性、稳定性方面的差异以及ART2wNF 在克服样本输入顺序影响等方面的优势.     

基于改进ART2算法的数据聚类方法

通过分析经典自适应谐振(adaptive resonance theory,ART)神经网络聚类过程中向量规格化、警戒参数全局化等特点,指出其无法应用于向量的模是与类别特征相关的情况、不能根据空间密度变化以不同粒度划分空间、输出无层次结构等不足,提出一种改进ART2神经网络算法.该算法提出警戒参数局部化和以模为标准的神经元预选,通过循环形成与模相关的多层次动态聚类结构(粗粒度的聚类不需重新训练神经网络),除此还降低对警戒参数主观设置的要求;这些优点能够适应模是重要特征的空间中子空间分别聚类的情况.     

气体传感器阵列常用模式识别算法

模式识别在气体传感器阵列的测量中占有举足轻重的地位。介绍了k近邻法、聚类分析、判别函数分析、反向传播人工神经网络、主元分析法、概率神经网、学习向量量化、自组织映射、自适应共振网、遗传算法等气体传感器阵列常用模式识别算法的原理和特点。同时,指出了在应用中模式识别算法选择和评价的标准。     

一种新的自适应谐振算法

本文提出了一个综合多种神经网络理论的学习算法ＦＴＡＲＴ（ｆｉｅｌｄｔｈｅｏｒｙ—ｂａｓｅｄａｄａｐｔｉｖｅｒｅｓｏｎａｎｃｅｔｈｅｏｒｙ），它将ＡＲＴ（ａｄａｐｔｉｖｅｒｅｓｏｎａｎｃｅｔｈｅｏｒｙ）算法、ＦｉｅｌｄＴｈｅｏｒｙ和ＡＲＴＭＡＰ等算法的优点有机结合，并以样本在实例空间出现的概率为启发信息修改分类．ＦＴＡＲＴ由于采用了不同于其它算法的冲突解决和动态扩大分类区域的方法，因此取得了较好的效果．本文还对实现ＦＴＡＲＴ算法的结果进行了验证．     

Adaptive resonance neural classifier for identification of gases/odours using an integrated sensor array

A new approach to intelligent gas sensor (IGS) design using a classifier based on adaptive resonance theory (ART) artificial neural network (ANN) is presented. Using published data of sensor arrays fabricated and characterised at our laboratory, we demonstrate excellent gas/odour identification performance of our classifier for a 4-gas, 4-sensor system to identify individual gas/odour. Since the ART neural network is a self-organising classifier trained in the unsupervised mode, it avoids the drawbacks associated with static feedforward neural networks trained with a locally optimal backpropagation-type training algorithms applied by researchers in the recent past. The ART neural network offers easy implementability and real time performance in addition to giving excellent classification accuracy as demonstrated by our experiments.     

Fuzzy lattice neural network (FLNN): a hybrid model for learning

This paper proposes two hierarchical schemes for learning, one for clustering and the other for classification problems. Both schemes can be implemented on a fuzzy lattice neural network (FLNN) architecture, to be introduced herein. The corresponding two learning models draw on adaptive resonance theory (ART) and min-max neurocomputing principles but their application domain is a mathematical lattice. Therefore they can handle more general types of data in addition to N-dimensional vectors. The FLNN neural model stems from a cross-fertilization of lattice theory and fuzzy set theory. Hence a novel theoretical foundation is introduced in this paper, that is the framework of fuzzy lattices or FL-framework, based on the concepts fuzzy lattice and inclusion measure. Sufficient conditions for the existence of an inclusion measure in a mathematical lattice are shown. The performance of the two FLNN schemes, that is for clustering and for classification, compares quite well with other methods and it is demonstrated by examples on various data sets including several benchmark data sets.     

一个新的在线学习文档分类模型

为了对在线学习文档进行分类,本文根据自适应谐振理论给出了一个半监督学习模糊ART模型(SLFART)及其算法.该算法不仅克服一般半监督学习算法中假设已知数据概率分布的条件,而且还具有非常强的学习新模式和纠正错误的能力.实验结果表明SLFART性能明显优于其它算法.     

基于神经网络的聚类方法研究

针对常用聚类方法不能有效处理噪声数据的问题,本文结合神经网络具有自适应性的特点,提出基于神经网络的聚类(NN_Cluster)模型,并设计了基于自适应共振理论的神经网络聚类模型(ARTNN_Cluster)和基于自组织特征映射的神经网络聚类模型(SOMNN_Cluster)。标准数据集上的实验结果表明,与传统的K_means聚类方法相比,本文提出的基于神经网络的聚类模型有效地克服了传统方法的噪声问题,得到了较好的聚类效果。     

一种自适应学习算法在图像识别中的应用研究

文章研究了特征图像的自适应识别技术问题,提出了利用自适应共振理论,实现自稳定和自组织识别的人工神经网络,达到学习和自适应识别图像的应用要求,根据这一原理及规则,给出了具有自稳定分类功能的自适应学习算法。     

二维空间聚类的树ART2模型

ART2网络是一种著名的聚类方法,已实际应用于诸多领域,其作用于二维空间数据,不仅存在模式漂移和向量幅度信息缺失的问题,而且难以适应不规则形态分布的空间数据的聚类。提出了一种树ART2网络模型(TART2),通过长期记忆(LTM)模式的调整和向量幅度信息的学习,使ART2网络保持了带空间距离约束的旧模式记忆;引入树结构优化,降低了警戒参数设置的主观要求,减少了模式交混现象的发生。对比实验结果表明,TART2网络更适用于带状分布的空间数据聚类,具有较高的可塑性和自适应性。     

Observations and problems applying ART2 for dynamic sensor patterninterpretation

This paper discusses characteristics of the ART2 (adaptive resonance theory) information processing model which emerge when applied to the problem of interpreting dynamic sensor data. Fast learn ART2 is employed in a supervised learning framework to classify process “fingerprints” generated from multi-sensor trend patterns. Interest in ART2 was motivated by the ability to provide closed classification regions, uniform hyperspherical clusters, feature extraction, and on-line adaption. Sensor data interpretation is briefly discussed with an emphasis on the unique attributes of the problem and the interaction with ART2 information processing principles. Pattern representations, e.g., time domain, which encode information in both magnitude and direction of the input vector are shown to be fundamentally incompatible with ART2. Complement coding is shown to solve this problem when the feature extraction capability of the ART2 network is disabled. Complement coding is also shown to preserve the clustering characteristics of the process “fingerprints” which are otherwise lost using the ART2 directional similarity measure. These issues are illustrated using an ART2-based monitoring system for a dynamically simulated chemical process