A prolog based expert system for the allocation of quality assurance program resources

The worldwide proliferation of fraudulent materials has brought about the need for a new approach to the control of purchased material quality. Nuclear power, defense, aerospace and many other industries are affected by the supply of poor quality clones that presume to be original replacement parts. Safety considerations abound as these highly defective materials are used in systems that are critical to the preservation of human lives. Commonly utilized quality assurance programs have not effectively stopped the influx of fraudulent materials. These programs fail to concentrate inspection efforts on materials and inspection attributes that would effectively prevent the acceptance of fraudulent materials. herein a solution is presented to this problem in the form of an expert system application. Information commonly available in industry is formulated into a knowledge based system wherein advisories are given to the user regarding key purchased material receipt inspection strategies.     

Application of Support Vector Machines in Assessing Conceptual Cost Estimates

Total conceptual cost estimates and the assessment of the quality of these estimates are critical in the early stages of a building construction project. In this study, the support vector machine (SVM) model for assessing the quality of conceptual cost estimates is proposed, and the application of SVM in construction areas is investigated. The results show that the SVM model assessed the quality of conceptual cost estimates slightly more accurately than the discriminant analysis model. This shows that using the SVM has potential in construction areas. In addition, the SVM model can assist clients in their evaluation of the quality of the estimated cost and the probability of exceeding the target cost, and in their decision on whether or not it is necessary to seek a more accurate estimate in the early stages of a project.     

基于奇异值分解和神经网络的个性化信息模式识别

随着信息资源的持续增长，信息获取系统为特定用户提供特定信息服务的需求越来越重要。该文提出了基于奇异值分解和神经网络的个性化信息模式识别的方法，可以方便地获取用户信息需求的特征，并且利用这些特征信息完成个性化的信息识别和匹配，即实现个性化信息的模式识别，进而实现信息的个性化获取和管理。该文中的算法实现简单，可以方便地应用到目前的一些信息管理、服务系统中。     

Determination of vitality from a non-invasive biomedical measurement for use in fingerprint scanners

Fingerprints are the oldest and most widely used biometrics for personal identification. Unfortunately, it is usually possible to deceive automatic fingerprint identification systems by presenting a well-duplicated synthetic or dismembered finger. This paper introduces one method to provide fingerprint vitality authentication in order to solve this problem. Detection of a perspiration pattern over the fingertip skin identifies the vitality of a fingerprint. Mapping the two-dimensional fingerprint images into one-dimensional signals, two ensembles of measures, namely static and dynamic measures, are derived for classification. Static patterns as well as temporal changes in dielectric mosaic structure of the skin, caused by perspiration, demonstrate themselves in these signals. Using these measures, this algorithm quantifies the sweating pattern and makes a final decision about vitality of the fingerprint by a neural network trained by examples.     

A neural network model for fault detection in conjunction with a programmable logic controller

This paper discusses the feasibility of using neural networks as a tool in the fault detection process. A neural network is integrated with a state language programmable logic controller, an important device in an automatic control system. Time series data related to time spent in a state is gathered and used as input into a neural network, for the purpose of identifying when a fault has occurred. A feedforward neural network is used to identify which (if any) of three types of faults may have occurred. Experimental results related to sensitivity and accuracy measures are presented. A brief review of related applications and research is also presented.     

神经网络分类器的快速构造算法

使用ＢＰ算法训练多层网络的速度很慢而且事先难于确定隐节点和隐层的适当数目。本文提出一个有效的算法，先构造决策树，然后将构造的决策树转换为神经网。文中使用一个全局准则函数控制决策树的增长，它较好地匹配了树的复杂性和训练样本量及错分率界。实验结果，本文的算法比用ＢＰ算法训练多层网络要快，而其分类精度不低于用ＢＰ算法训练的多层神经网。     

实现CDMA最佳多用户检测的神经网络方法

提出并讨论了两种实现码分多址系统中最佳多用户检测的神经网络方法。一种基于离散Ｈｏｐｆｉｅｌｄ神经网络，另一种基于采用反向传播算法的多层感知器神经网络。理论分析和计算机模拟都证实了这两种神经网络方法的可行性，优越性和实用性。前者适用“固定”用户情况；后者既可用于“固定”用户吼适用于移动用户的情况，因而在未来的ＣＤＭＡ个人通信网中有较好的应用前景。     

基于神经PID网络的串级调节系统研究

本文提出了一种神经PID网络控制器，并利用它构成串级调节系统。在给出网络结构和计算公式的基础上，进行了计算机仿真。仿真结果表明，神经PID网络控制器构成的串级调节系统具有良好的动态和静态性能及很强的抗扰能力。     

基于TMN的接入网网管系统模型

本文介绍了 TMN 网管系统的体系结构,进而讨论了接入网网管系统的体系结构和管理接口的有关问题。     

Neural networks,rational functions,and realization theory

The problem of parametrizing single hidden layer scalar neural networks with continuous activation functions is investigated. A connection is drawn between realization theory for linear dynamical systems, rational functions, and neural networks that appears to be new. A result of this connection is a general parametrization of such neural networks in terms of strictly proper rational functions. Some existence and uniqueness results are derived. Jordan decompositions are developed, which show how the general form can be expressed in terms of a sum of canonical second order sections. The parametrization may be useful for studying learning algorithms.This work was supported by the Australian Research Council, the Australian Telecommunications and Electronics Research Board, and the Boeing Commencai Aircraft Company (thanks to John Moore).