关于人工神经网络

本文对神经网络的基本原理和应用进行了深入浅出的阐述,首先说明为何引进神经网络,然后讲述神经网络的基本原理和当前有关神经网络的前沿技术,其后对神经网络的应用情况进行了介绍,最后对神经网络有待解决的问题进行了简要说明。     

人工神经网络在火灾探测中的应用

火灾的早期探测是较为复杂且具有重要意义的研究课题。针对传统火灾探测方法存在的不足,提出了一种基于人工神经网络技术的火灾自动探测系统,即采用黄金分割优选法对网络隐含层节点数进行优选,采用优于BP算法的LM算法进行网络学习训练。实验结果表明不仅这种系统具有拟合精度高和运算速度快等优点,且探测效率良好。     

人工神经网络及其在分析化学中的应用

讨论了人工神经网络的基本原理，对神经网络在分析化学领域的应用作了综述。     

人工神经网络在工业机器人中的应用

Artificial neural networks (ANNs) have been widely used to solve a number of problems to which analytical solutions are difficult to obtain using traditional mathematical approaches. Such problems exist also in the analysis of industrial robots. This paper presents an overview of ANN applications to robot kinematics, dynamics, control, trajectory and path planning, and sensing. Reasons for using or not using ANNs to industrial robots are explained as well.     

基于人工神经网络的宏观经济预测模型

运用人工神经网络方法,提出了将神经网络用于宏观经济预测中的一般步骤和方法,并以我国的国内生产总值为例进行宏观经济预测建模,为经济预测提供了新的思想和途径.     

BP神经网络研究及其地学应用

误差反向传播学习网络是目前运用最广、实现途径最直观的一种神经网络,而且是已知的非线性映射能力最强的一种神经网络模型之一。它在模式识别、知识发现、模式分类等方面具有很强的功能。在地学研究中,误差反向传播学习网络模型可以很好地模拟致矿地质因素与成矿的复杂非线形映射关系。     

Artificial Neural Network Model for Optical Fiber Direction Coupler Design

A new approach to the design of the optical fiber direction coupler by using neural network is proposed. To train the artificial neural network, the coupling length is defined as the input sample, and the coupling ratio is defined as the output sample. Compared with the numerical value calculation of the theoretical formula, the error of the neural network model output is 1% less. Then, through the model, to design a broadband or a single wavelength optical fiber direction coupler becomes easy. The method is proved to be reliable, accurate and time-saving. So it is promising in the field of both investigation and application.     

基于人工神经网络的个人信用评估模型

针对个人信用评估，提出了人工神经网络的评估模型。对于训练好的网络模型，用户可在输入客户数据的基础上，调入生成好的规则，正确得到客户的评价结果。     

神经网络在电力中期负荷预测中的应用

为准确预测电力系统中期负荷，针对常用BP算法的预测速度慢、易陷入局部最优解的缺点，改进了基本BP算法，建立了中期负荷预报模型。负荷预测仿真表明，对BP算法的改进可使预测精度和收敛速度均得到显著的提高．     

Condition Monitoring and Diagnostics for Rotating Machinery Using Artificial Neural Networks

ＣｏｎｄｉｔｉｏｎＭｏｎｉｔｏｒｉｎｇａｎｄＤｉａｇｎｏｓｔｉｃｓｆｏｒＲｏｔａｔｉｎｇＭａｃｈｉｎｅｒｙＵｓｉｎｇＡｒｔｉｆｉｃｉａｌＮｅｕｒａｌＮｅｔｗｏｒｋｓＹａｎｇＺｈｏｎｇ（杨忠）ＳｈｅｎｇＺｈａｏｈａｎ（盛昭瀚）（Ｉｎｓｔｉｔｕｔｅｏｆ...     

Visualization of Twitter Data in the Classroom

The expression “big data” is ubiquitous in the business world today, but few undergraduate business students have the opportunity to gain practical experience with how new business analytics tools can be used in decision making. This article describes a set of hands‐on labs that prepare students to incorporate streaming data analysis into group research projects. Splunk is used to help students analyze and visualize streaming social media data. An evaluation of student projects and student survey results show that this practical approach of training students to manipulate and visualize big data was largely successful in achieving instructional goals.     

Creating Impactful Student Learning in a Business Analytics Program through Leveraging Scholar‐Practitioner Engagement

Industry demand for talented analytics professionals has created a significant increase in degree programs (e.g., MS in Analytics) around the globe. Many of these programs incorporate experiential learning in the curriculum to foster a deeper understanding. This article focuses on the value and challenges in implementing experiential learning in an analytics‐focused degree program by incorporating and scaffolding multiple organizational analytics projects throughout the curriculum. In addition, this article focuses on the manner in which these organizational analytics projects can create value for scholars, beyond student learning. Scholar‐practitioner partnerships have the potential to advance not only the field of academic research but also the rigor of actual practice. We report our experiences and best practices in creating and leveraging scholar‐practitioner partnerships in the context of project‐based experiential learning in a Master of Science in Business Analytics (MSBA) program. We provide motivation for creating such partnerships for teaching analytics, document student and organizational success, offer illustrations of exemplary experiential learning projects, and discuss the challenges that need to be managed.     

人工神经网络在教育资源管理中的应用

随着教育资源数据量、内容和形式的极大丰富,教育资源的组织和管理日益成为人们关注的焦点。文章指出了传统教育资源管理系统在教育资源飞速增长的情况下所存在的一系列问题,然后以此为基础,提出了将人工神经网络应用于教育资源管理中的初步设想。同时结合神经网络的结构特点,设计了教育资源管理系统模型,构建了教育资源管理的新模式,探讨了其在教育资源管理中所发挥的优势。     

人工神经网络电路的新发展

简要回顾人工神经网络电路的发展,讨论两种应用最为广泛的模型,介绍目前人工神经网络电路研究的最新进展.     

基于人工神经网络的无机抗菌剂抗菌性能预测模型

Quantitatively evaluation of antibacterial activities of inorganic antibacterial agents is an urgent problem to be solved.Using experimental data by an orthogonal design, a prediction model of the relation between conditions of preparing inorganic antibacterial agents and their antibacterial activities has been developed.This is accomplished by introducing BP artificial neural networks in the study of inorganic antibacterial agents.It provides a theoretical support for the development and research on inorganic antibacterial agents.     

基于BP神经网络的网络学习评价模型研究

研究网络学习评价问题对推动网络教学资源的使用具有十分重要的意义,传统的网络学习评价方法具有很强的主观性,且仅限于线性模型,缺少科学性。为克服传统网络学习评价方法的不足,实现网络学习评价的智能化,提出了一种基于BP神经网络的的网络学习评价模型,并利用MATLAB进行实验仿真,测试结果表明,该评价模型准确率高,能为网络学习评价提供可靠数据。     

综合素质评价的神经网络模型及Simulink实现

针对大学生综合素质评价过程中存在诸多非定量因素及非线性特征,建立综合素质测评的BP神经网络模型。利用BP神经网络实现对量化评价函数的逼近。所给出的神经网络模型较已有的评价体系具有更高的执行效率,而且还能生成模块化的程序进行推广。利用Matlab软件的Simulink平台对BP神经网络进行仿真计算,该方法可操作性强,不需要进行模糊评测方法过程中的复杂运算,具有较广泛的实用性。     

基于神经网络的数据挖掘算法

数据挖掘涉及多学科技术的集成,包括数据库技术、神经网络、模式识别、信息检索、图像与信号处理和空间数据分析等.神经网络在数据挖掘中的应用也越来越广泛,尽管神经网络具有结构复杂、网络训练时间长、结果表示不容易理解等缺陷,但其对噪声数据的高承受能力和低错误率的优点是其他方法所不及的,并在数据挖掘所采用的方法中具有优势.本文对基于BP神经网络的数据挖掘算法进行了详细的讨论,并提出了算法改进方法.     

对一离散Hopfield神经网络稳定性定理的补充

对一离散型Hopfield神经网络的稳定性定理进行讨论,对其结论作了补充修改,并以具体例子说明了定理修改的必要性.     

基于人工神经网络的模式识别

利用数学软件MATLAB研究了基于人工神经网络的模式识别,这种基于人工神经网络的模式识别系统不仅能够识别字母还能识别汉字和数字,适用于汉字识别和数字签名。