基于Levenberg-Marquardt神经网络的企业资信评估方法研究

企业资信评估问题是一个复杂的非线性问题,而神经网络技术可实现非线性关系的隐式表达。文章提出将基于Levenberg-Marquardt算法的多层前馈型神经网络用于资信评估,并通过MATLAB软件及其神经网络工具对其进行仿真计算。实验结果表明,企业资信神经网络评估模型收敛速度快,准确率较高,具有一定的实用价值。     

FTSET-a software tool for fault tolerance evaluation and improvement

It is commonly assumed that neural networks have a built-in fault tolerance property mainly due to their parallel structure. The international community of neural networks discussed these properties until 1994 and afterward the subject has been mostly ignored. Recently, the subject was again brought to discussion due to the possibility of using neural networks in areas where fault tolerance and graceful degradation properties would be an added value, like medical applications of nano-electronics or space missions. Nevertheless, the evaluation of fault tolerance and graceful degradation characteristics remained difficult because there were no systematic methods or tools that could be easily applied to a given Artificial Neural Networks application. The discussion of models is the first step for sorting ways of developing the fault tolerance capability and for building a tool that can evaluate and improve this characteristic. The present work proposes a fault tolerance model, presents solutions for improving it and introduces the Fault Tolerance Simulation and Evaluation Tool for Artificial Neural Networks that evaluates and improves fault tolerance.     

API-Evolution Support with Diff-CatchUp

Applications built on reusable component frameworks are subject to two independent, and potentially conflicting, evolution processes. The application evolves in response to the specific requirements and desired qualities of the application's stakeholders. On the other hand, the evolution of the component framework is driven by the need to improve the framework functionality and quality while maintaining its generality. Thus, changes to the component framework frequently change its API on which its client applications rely and, as a result, these applications break. To date, there has been some work aimed at supporting the migration of client applications to newer versions of their underlying frameworks, but it usually requires that the framework developers do additional work for that purpose or that the application developers use the same tools as the framework developers. In this paper, we discuss our approach to tackle the API-evolution problem in the context of reuse-based software development, which automatically recognizes the API changes of the reused framework and proposes plausible replacements to the "obsolete" API based on working examples of the framework code base. This approach has been implemented in the Diff-CatchUp tool. We report on two case studies that we have conducted to evaluate the effectiveness of our approach with its Diff-CatchUp prototype.     

Object-oriented modeling and simulation of neural networks

The field of neural networks is being investigated by many researchers in order to provide solutions to difficult problems in the area of manufacturing systems. Computer simulation of neural networks is an important part of this investigation. This paper applies concepts from an important trend in software engineering research, namely object-oriented programming, to model neural networks.The design and implementation of a software object library is crucial to obtaining the full benefits of object-oriented programming. In this paper we discuss the design and implementation of a foundation library of software objects for the purpose of simulating and validating different network architectures and learning rules. The library contains objects that implement various types of nodes and learning rules. We discuss the results of our experiments to illustrate the benefits of using an object-oriented approach to modeling neural networks.     

无线传感器网络中基于RSSI的节点距离预测

无线传感器网络大量应用在环境监测、目标跟踪、安全监控等领域,因此网络的自身定位是大多数应用的基础。常用的定位方法必须测量节点间的距离。为了预测距离值,根据实验获取的RSSI值与对应的距离值,先对实验数据进行滤波处理,建立面向Matlab神经网络工具箱的神经网络预测模型,利用神经网络的特性和Matlab工具箱的强大功能,通过实测数据对网络进行训练。预测结果表明,距离精度达到1 m之内。     

径向基函数网络在数值计算中的应用

由经典的函数逼近理论衍生的很多数值算法有共同的缺点：计算量大、适应性差、对模型和数据要求高,在实际应用中受到限制。神经网络可以被用来计算复杂输入与输出结果之间的关系,故神经网络具有很强的函数逼近功能。该文给出了径向基函数网络（RBFNN）的结构及学习过程,重点阐述了RBFNN在函数逼近、求解非线性方程组以及散乱数据插值中的应用,结合MATLAB神经网络工具箱给出了数值实例,并与BP网络进行了比较。应用结果表明RBFNN是数值计算的一个有力工具,与传统方法比较具有编程简单、实用的特点,在工程和科学研究上若将其制成软件包则具有很好的使用价值。     

基于神经网络的数模混合电路故障诊断模型设计

神经网络是智能故障诊断中运用最为广泛的诊断方法之一,能应对复杂数字电路和模拟电路故障诊断;基于VC++的神经网络诊断系统具有运算速度快、输入输出界面丰富、易于实现在线故障诊断等特点,而Matlab提供强大的神经网络工具箱,在网络训练方面有优势;利用参数文件将二者优势结合起来,可实现功能完善的数模混合电路智能故障诊断系统;诊断实例表明,使用该方法实现数模混合电路故障诊断系统,具有高建模速度和高诊断精度的特点。     

A component framework for reusing a proprietary computer-aided engineering environment

Nowadays, it is becoming more frequent for engineers to deal with problems and situations that require specific software and the commercially available applications may not result adequate. Because of this, the reuse of software components is becoming a normal practice for improving the productivity of the software programmers, and the quality of the products that they develop. The employment of reusable components presents a number of advantages, such as the reduction in time necessary to develop new software, or the simplification of many computational tasks.There are some proprietary software environments in the engineering domain that are practically de facto standards, since they offer a broad functionality, robustness and constant updating. Therefore, those environments could be ideal candidates to be reusable components when developing new software. An example of this is Matlab, which can be encapsulated, in order to use it as a true reusable component. This will provide the user with efficient tools for designing his/her own applications.Moreover, there is an increasing importance for any software to offer distributed services. To perform this task, it is fundamental to have at one’s disposal reusable components that support Internet-based distributed applications and services development.In the present work, a software component framework that effectively encapsulates Matlab is described. This software allows developers the reuse of Matlab, for both local and distributed applications. To address those issues the .NET technology was employed. The component framework developed can be integrated in the toolbars of software development environments supporting the .NET framework. This facilitates the construction of applications that can reuse the components, since the software developers can use and test them and change their properties in design-time.     

BP神经网络及其在数据分类中的应用

BP神经网络模型是一种典型的前向型神经网络,具有良好的自学习、自适应、联想记忆、并行处理和非线形转换的能力,是目前应用最为广泛的一种神经网络模型。本文介绍了BP神经网络的实现以及其在数据挖掘分类方面的应用。     

传感器故障检测的仿真研究

该文主要介绍了连续小波变换和BP神经网络在传感器故障诊断中的应用,并就如何将小波变换和神经网络结合起来进行故障检测进行了有益的探讨、研究,最终借助于matlab软件和它自带的simulink工具以及相关的信号处理工具包做了仿真算例,获得了满意的效果,证明了这两种方法的可行性。     

Pulse coupled neural networks and its applications

This paper surveys the extensive usage of pulse coupled neural networks. The visual cortex system of mammalians was the backbone for the development of pulse coupled neural network. PCNN (Pulse Coupled Neural Networks) is unique from other techniques due to its synchronous pulsed output, adjustable threshold and controllable parameters. is Hence the uniqueness of this network utilized in the fields of image processing. The basic model of PCNN and the consecutive changes implemented, to strengthen the pulse coupled neural network are discussed initially. Then the applications of PCNN are broadly discussed. The other miscellaneous applications utilizing pulse coupled neural networks are thrown light in the last section.     

以项目为中心的面向对象复用支持

现有的软件复用技术通常是围绕着库来组织利用标准的和通用的可复用资源.例如,面向对象编程环境中的类库和通用构件库.然而,这种以库为中心的复用方式在一定程度上忽略了可复用资源的项目相关信息,而项目相关信息记录了可复用资源的应用语境(Application Context).应用语境有利于可复用资源的理解和使用.为此,提出了一种围绕着一个项目的文档来组织和利用可复用资源的复用途径,即以项目为中心的文档复用,用以支持在相同应用领域中一族软件的开发.探讨了当前的面向对象方法在支持文档复用方面的一些局限性,提出了一种     

Componentization: the Visitor example

Patterns provide a vocabulary and catalog for common design solutions but are not reusable software. Through advanced object-oriented language mechanisms it is possible to simplify the application developer's job by turning certain patterns into fully reusable solutions.     

Object-oriented software fault prediction using neural networks

This paper introduces two neural network based software fault prediction models using Object-Oriented metrics. They are empirically validated using a data set collected from the software modules developed by the graduate students of our academic institution. The results are compared with two statistical models using five quality attributes and found that neural networks do better. Among the two neural networks, Probabilistic Neural Networks outperform in predicting the fault proneness of the Object-Oriented modules developed.     

基于贝叶斯正则化网络的绿色创新战略选择

科学、准确地进行企业绿色技术创新战略选择对环境保护和企业长期持续发展具有重要意义。文章提出将基于贝叶斯正则化神经网络用于企业绿色技术创新战略选择,并通过MATLAB神经网络工具对其进行模拟计算。实验结果表明了该方法的有效性。     

Speech recognition with artificial neural networks

In this paper, artificial neural networks were used to accomplish isolated speech recognition. The topic was investigated in two steps, consisting of the pre-processing part with Digital Signal Processing (DSP) techniques and the post-processing part with Artificial Neural Networks (ANN). These two parts were briefly explained and speech recognizers using different ANN architectures were implemented on Matlab. Three different neural network models; multi layer back propagation, Elman and probabilistic neural networks were designed. Performance comparisons with similar studies found in the related literature indicated that our proposed ANN structures yield satisfactory results.     

基于C# .NET与Matlab接口和BP网络的汽车产量预测

为了提高软件预测系统的开发效率和精度,研究了误差反向传播神经网络(Error Back Propagation Neural Network,BPNN)在Matlab 2006a环境下的实现方法,采用迭代法选择神经网络训练参数。探讨了Matlab与C# .NET开发平台的接口技术,能成功实现C# .NET对Matlab神经网络工具箱的调用。以汽车产量预测为应用案例,开发了基于C# .NET和Matlab接口的神经网络预测系统。实验结果表明,该预测系统对汽车产量预测有着良好性能,接口技术的应用提高了软件开发效率。     

粗糙集和主元分析在基于神经网络的模拟电路故障诊断中的应用

针对基于神经网络的模拟电路故障诊断中,故障特征集维数过高带来的诊断难点,提出了利用粗糙集和主元分析法对故障特征集进行预处理。粗糙集对故障诊断决策表进行属性约简,主元分析进行数据压缩及特征提取。试验仿真表明,对预处理后的数据进行识别,简化了神经网络结构,可有效提高网络的训练速度与诊断效率。     

设计模式及其在AMCCS中的应用

虽然目前已存在多种面向对象的软件设计方法,但在面向对象建模中仍存在不尽人意之处。进行面向对象的软件设计是比较困难的,而设计可复用的面向对象软件就更加困难。设计模式是面向对象建模的一种辅助设计手段,它总结了面向对象设计的成功设计经验,使得在遇到相似场景时可以直接套用。结合民航维修成本控制系统(AMCCS)的开发工作,使用设计模式细化具体组件和部件的对象建模方法,取得了较好的效果。     

Benchmarking library and application software with Data Envelopment Analysis

Library software is generally believed to be well-structured and follows certain design guidelines due to the need of continuous evolution and stability of the respective APIs. We perform an empirical study to investigate whether the design of open-source library software is actually superior to that of application software. By analyzing certain design principles and heuristics that are considered important for API design, we extract a set of software metrics that are expected to reflect the improved nature of libraries. An initial comparison by conventional statistical analysis confirms the overall belief that products of different software size scale should not be compared by simply examining metric values in isolation. In this paper, we propose the use of Data Envelopment Analysis (DEA), borrowed from production economics, as a means of measuring and benchmarking the quality of different object-oriented software designs captured by software metrics and apply this approach to the comparison of library and application software. The advantages offered by DEA and the differences between the application of DEA in an economic and a technological context are discussed. Results of the approach are presented for 44 open-source projects, equally divided between libraries and applications.