检索结果-内蒙古大学图书馆

Estimation of simultaneous equation models by backpropagation method using stochastic gradient descent

PEERJ COMPUTER SCIENCE 2024年 10卷 e2352-e2352页

作者： Perez-Sanchez, Belen Perea, Carmen Ballester, Guillem Duran Lopez-Espin, Jose J. Univ Miguel Hernandez Elche Ctr Operat Res Elche Alicante Spain Fragile Tech Palma De Mallorca Mallorca Spain

Simultaneous equation model (SEM) is an econometric technique traditionally used in economics but with many applications in other sciences. This model allows the bidirectional relationship between variables and a simultaneous relationship between the equation set. There are many estimators used for solving an SEM. Two-steps least squares (2SLS), three-steps least squares (3SLS), indirect least squares (ILS), etc. . are some of the most used of them. These estimators let us obtain a value of the coefficient of an SEM showing the relationship between the variables. There are different works to study and compare the estimators of an SEM comparing the error in the prediction of the data, the computational cost, etc. . Some of these works study the estimators from different paradigms such as classical statistics, Bayesian statistics, non-linear regression models, etc. . This work proposes to assume an SEM as a particular case of an artificial neural networks (ANN), considering the neurons of the ANN as the variables of the SEM and the weight of the connections of the neurons the coefficients of the SEM. Thus, backpropagation method using stochastic gradient descent (SGD) is proposed and studied as a new method to obtain the coefficient of an SEM.

关键词： backpropagation method Stochastic gradient descent Simultaneous equation models Artificial neural networks

来源：评论

学校读者我要写书评

暂无评论

A neural network-based method for coverage measurement of shot-peened panels

引用

NEURAL COMPUTING & APPLICATIONS 2019年第9期31卷 4829-4836页

作者： Shahid, Lubna Janabi-Sharifi, Farrokh Ryerson Univ Mech & Ind Engn Dept RMAL Robot Mechatron & Automat Lab 87 Gerrard St East Toronto ON M5B 2K3 Canada

Shot peening is a cold metal working process to improve the material strength, reduce corrosion fatigue and prevent fracture. Measuring the coverage level is an essential parameter in shot peening, which is traditionally performed through manual visual inspection. Due to the tedious nature of the task, it is prone to imprecision caused by human error. Several image processing and computer vision techniques are proposed in the literature to automate this process. While most of the techniques are accurate in segmenting the shot-peened areas, they seem to fail in the presence of machining streaks, resulting in false segmentation. To overcome this challenge, an artificial neural network (ANN)-based implementation is employed in this paper to improve accuracy of the results. The neural network is trained with specific selected features from the acquired images. Results show ANN outperforms the previously implemented standard image segmentation methods.

关键词： Image segmentation Shot peen coverage Vision-based coverage measurement Artificial neural networks backpropagation method Machining streaks

来源：评论

学校读者我要写书评

暂无评论

Field-programmable gate array-based functional link radial basis function network control for permanent magnet linear synchronous motor servo drive system

引用

IET ELECTRIC POWER APPLICATIONS 2010年第5期4卷 357-372页

作者： Lin, F. -J. Chou, P. -H. Hung, Y. -C. Wang, W. -M. Natl Cent Univ Dept Elect Engn Chungli 320 Taiwan Natl Dong Hwa Univ Dept Elect Engn Hualien 974 Taiwan

A field-programmable gate array (FPGA)-based functional link radial basis function network (FLRBFN) control is proposed in this study to control the mover of a permanent magnet linear synchronous motor (PMLSM) servo drive system to track periodic reference trajectories. First, the dynamics of the field-oriented control PMLSM servo drive with a lumped uncertainty, which contains parameter variations, external disturbances and non-linear friction force, is derived. Then, to achieve accurate trajectory tracking performance with robustness, an intelligent control approach using FLRBFN is proposed for the field-oriented control PMLSM servo drive system. The proposed FLRBFN is a radial basis function network (RBFN) embedded with a functional link neural network (FLNN). Moreover, the on-line learning algorithm of the FLRBFN, including the connective weights, the centres and the centres' width of the receptive field functions, are derived using back-propagation (BP) method. Furthermore, an FPGA chip is adopted to implement the developed control and on-line learning algorithms for possible low-cost and high-performance industrial applications using PMLSM. Finally, the effectiveness of the proposed control scheme and the robustness to parameter variations, external disturbances and friction force of the PMLSM servo drive system are verified by some experimental results.

关键词： field-oriented control linear motors machine vector control Control of electric power systems backpropagation method lumped uncertainty FPGA neurocontrollers permanent magnet linear synchronous motor servo drive system synchronous motor drives Linear machines Actuating and final control devices trajectory tracking performance RBFN functional link radial basis function network control field programmable gate arrays servomechanisms permanent magnet motors Synchronous machines Neurocontrol field-programmable gate array backpropagation radial basis function networks nonlinear friction force intelligent control approach Drives

来源：评论

学校读者我要写书评

暂无评论

MODEL-FREE FORECASTING FOR NONLINEAR TIME-SERIES (WITH APPLICATION TO EXCHANGE-RATES)

引用

COMPUTATIONAL STATISTICS & DATA ANALYSIS 1995年第4期19卷 433-459页

作者： WU, B NATL CHENGCHI UNIV INST SCI MAT DEPT MATH SCI TAIPEI TAIWAN

The first step in nonlinear model construction is data identification. While the major drawback with most of the nonlinear tests is that they are designed to test a parametric null against certain parametric alternatives and are therefore inconsistent against all possibile alternatives. In this paper, by incorporating the neurocomputing technique, the intriguing model construction and forecast problems such as how to find a best fitted model and to obtain a more appropriate prediction performance are addressed. The dynamic abilities of this approach lies in the fact that the multilier feed forward networks are functional approximators for any functions. Simulations verify that the neural networks perform a robust forecast for some nonlinear time series. Finally, we use the data of the exchange rates of Taiwan to compare the predictive performances of neural networks with ARIMA models.

关键词： MODEL-FREE FORECASTING NEURAL NETWORKS backpropagation method BILINEAR MODELS EXCHANGE RATES

来源：评论

学校读者我要写书评

暂无评论

Adaptive recurrent fuzzy neural network control for synchronous reluctance motor servo drive

引用

IEE PROCEEDINGS-ELECTRIC POWER APPLICATIONS 2004年第6期151卷 711-724页

作者： Lin, CH Natl United Univ Dept Elect Engn Miaoli 360 Taiwan

In the paper an adaptive recurrent fuzzy neural network (ARFNN) control system is proposed, to control a synchronous reluctance motor (SynRM) servo drive. First, the field-oriented mechanism is applied to formulate the dynamic equation of the SynRM servo drive. Then, the ARFNN control system is proposed to control the rotor of the SynRM servo drive for the tracking of periodic reference inputs. In the ARFNN control system, the RFNN controller is used to mimic an optimal control law, and the compensated controller with adaptive algorithm is proposed to compensate for the difference between the optimal control law and the RFNN controller. Moreover, an online parameter training methodology, which is derived using the Lyapunov stability theorem and the backpropagation method, is proposed to increase the learning capability of the RFNN. The effectiveness of the proposed control scheme is verified by simulated and experimental results.

关键词： synchronous reluctance motor servo drive control machine vector control online parameter training methodology reluctance motor drives adaptive algorithm adaptive control fuzzy control Lyapunov stability theorem Power engineering computing fuzzy neural nets optimal control Optimal control backpropagation method Lyapunov methods servomotors compensated controller rotors Neural computing techniques Synchronous machines field-oriented mechanism rotor tracking backpropagation Small and special purpose electric machines Fuzzy control adaptive recurrent fuzzy neural network Drives Control of electric power systems Self-adjusting control systems recurrent neural nets optimal control law Stability in control theory

来源：评论

学校读者我要写书评

暂无评论

Adaptive hybrid control using a recurrent neural network for a linear synchronous motor servo-drive system

引用

IEE PROCEEDINGS-CONTROL THEORY AND APPLICATIONS 2001年第2期148卷 156-168页

作者： Lin, CH Chou, WD Lin, FJ Natl Lien Ho Inst Technol Dept Elect Engn Miaoli 360 Taiwan Chung Yuan Christian Univ Dept Elect Engn Chungli 32023 Taiwan

An adaptive hybrid control system using a recurrent neural network (RNN) is proposed to control a permanent magnet linear synchronous motor (PMLSM) servo-drive system. First, a field-oriented mechanism is applied to formulate the dynamic equation of the PMLSM servo-drive. Then, a hybrid control system is proposed to control the mover of the PMLSM servo-drive for periodic motion. In the hybrid control system, the RNN controller is the main tracking controller, which is used to mimic an optimal control law, and the compensated controller is proposed to compensate the difference between the optimal control law and the RNN controller. Moreover, an on-line parameter training methodology of the RNN, which is derived using the Lyapunov stability theorem and the backpropagation method is proposed to guarantee the asymptotic stability of the control system. In addition, to relax the requirement for the bounds of minimum approximation error and Taylor high-order terms, an adaptive hybrid control system is investigated to control the PMLSM servo-drive, where two simple adaptive algorithms are utilised to estimate the mentioned bounds. The effectiveness of the proposed control schemes is verified by both the simulated and experimental results.

关键词： Electric actuators and final control equipment Learning in AI online parameter training methodology asymptotic stability adaptive control adaptive hybrid control system PM motor Taylor high-order terms Lyapunov stability theorem Adaptive system theory periodic motion permanent magnet linear synchronous motor linear synchronous motors synchros dynamic equation optimal control permanent magnet motors backpropagation method Neural nets Lyapunov methods PMLSM servodrive Neurocontrol Synchronous machines machine control field-oriented mechanism recurrent neural network compensation RNN backpropagation learning (artificial intelligence) neurocontrollers linear synchronous motor servo-drive system Control of electric power systems Drives Self-adjusting control systems minimum approximation error bounds recurrent neural nets Stability in control theory synchronous motor drives

来源：评论

学校读者我要写书评

暂无评论

Deep adversarial reconstruction classification network for unsupervised domain adaptation

引用

INTERNATIONAL JOURNAL OF MACHINE LEARNING AND CYBERNETICS 2024年第6期15卷 2367-2382页

作者： Lin, Jiawei Bian, Zekang Wang, Shitong Jiangnan Univ Sch Artificial Intelligence & Comp Sci Wuxi 214122 Jiangsu Peoples R China

Although the existing adversarial domain adaptation methods have been successfully applied in the unsupervised domain adaptation community, their performances may perhaps be weakened due to a significant distributional diversity of the target domain caused by the absence of the local features of samples in the target domain after the domain adaptation process. In this paper, based on the well-known autoencoder, a single-input with multi-output model called deep adversarial reconstruction classification network (DARCN) is developed to circumvent the above issue. The proposed DARCN mainly consists of the following four modules: a feature extractor for extracting the domain-invariant features along with the local features of the target domain, a predictor for estimating the label, a domain discriminator for distinguishing between the source and target domains, and a decoder for reconstructing the original data. The standard backpropagation method can be effectively used to optimize the proposed model. The experimental results indicate that DARCN realizes the improved classification performances in most cases in contrast to some existing comparative methods.

关键词： Unsupervised domain adaptation Local features Autoencoder backpropagation method

来源：评论

学校读者我要写书评

暂无评论

Designing fuzzy logic systems

引用

IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS 1997年第11期44卷 885-895页

作者： Mendel, JM Mouzouris, GC Signal and Image Processing Institute Department of Electrical Engineering-Systems University of Southern California Los Angeles CA USA

We present a formulation of a fuzzy logic system (FLS) that can be used to construct nonparametric models of nonlinear processes, given only input-output data. In order to effectively construct such models, we discuss several design methods with different properties and features. We compare and illustrate systems designed with each one of the methods, using an example on the predictive modeling of a nonlinear dynamic (chaotic) system.

关键词： backpropagation method chaotic time series design forecasting fuzzy logic systems one-pass method RLS method

来源：评论

学校读者我要写书评

暂无评论

AN ADAPTIVE MOMENTUM BACK-PROPAGATION (AMBP)

引用

NEURAL COMPUTING & APPLICATIONS 1995年第4期3卷 213-221页

作者： DRAGO, GP MORANDO, M RIDELLA, S UNIV GENOA DIBE I-16145 GENOA ITALY

An algorithm for fast minimum search is proposed, which achieves very satisfying performance harmonising the Vogl's and the Conjugate Gradient algorithms. Such effectiveness is achieved by making adaptive, in a very simple and satisfactory way, both the rate and the term, and by momentum and corrections both on the possible cost function increase and on moves opposite to the direction of the negative of the gradient. Thanks to these improvements, we can obtain a good scaling relationship in learning. As regards the real world context, a musical application showed favourable results: besides the good convergence speed, a high generalisation capability has been achieved, as confirmed both by subjective musical evaluations and by objective tests.

关键词： backpropagation method GENERALIZATION TEST LEARNING MULTILAYERED PERCEPTRON MUSICAL COMPOSITION NEURAL NETWORK COMPUTING

来源：评论

学校读者我要写书评

暂无评论

EFFICIENT NEURAL-NETWORK TRAINING ON A CRAY Y-MP

引用

INTERNATIONAL JOURNAL OF HIGH SPEED COMPUTING 1995年第1期7卷 109-123页

作者： CHUNG, SL SETIONO, R Department of Information Systems and Computer Science National University of Singapore Kent Ridge Singapore 0511 Republic of Singapore

An efficient implementation of a quasi-Newton algorithm for training feed-forward neural network on a Gray Y-MP is presented. The most time-consuming step of a neural network training using the quasi-Newton algorithm is the computation of the error function and its gradient. Parallelization embedded in these computations can be exploited through vectorization in a Gray Y-MP supercomputer. We show how they can be carried out such that the overall performance of the neural network training process can be enhanced substantially.

关键词： NEURAL NETWORK TRAINING QUASI-NEWTON methodS backpropagation method VECTORIZATION

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：