检索结果-内蒙古大学图书馆

2010 IEEE World Congress on Computational Intelligence

作者： Yeh, Jen-Wei Su, Shun-Feng Jeng, Jin-Tsong Chen, Bor-Sen Natl Taiwan Univ Sci & Technol Dept Elect Engn Taipei Taiwan Natl Formosa Univ Dept Informat Engn Fuwei Taiwan Natl Tsing Hua Univ Dept Elect Engn Hsinchu Taiwan

ISBN: (纸本)9781424469208

Self-constructing neural fuzzy inference network (SONFIN) is a neural fuzzy system and owing to its structure learning capability, SONFIN has been demonstrated to have excellent learning performance. However, various parameters must be selected in the implementation of SONFIN. In this paper, the learning behavior of SONFIN is studied. First, the SONFIN system with different thresholds and variances are considered. Different selections will result in different rule numbers and membership function widths. Our experiment results indicate that when there are possibilities of overfitting, more rules may not always come up with better performance. Secondly, two different learning algorithms are considered;the backpropagation (BP) learning algorithm and the recursive least square (RLS) algorithm. It can found that the learning of using RLS is much faster than that of using BP as expected. However, it can be found that when overfitting may occur, BP can have better learning performance in terms of testing errors. Finally, the use of reset for the covariance matrix in the RLS algorithm is investigated. From this primitive study, it can be found that the learning algorithms or some parameter selections may have both good effects in the testing performance and the training performance before the learning does not have significant overfitting. However, when the learning crosses this point, any selection is good for learning may bring bad effects on the testing performance.

关键词： RLS algorithm SONFIN system backpropagation backpropagation learning algorithm covariance matrices covariance matrix fuzzy neural nets fuzzy reasoning learning analysis learning behavior learning performance least squares approximations membership function neural fuzzy system parameter selection recursive least square algorithm self-constructing neural fuzzy inference network testing performance training performance

来源：评论

学校读者我要写书评

暂无评论

Analysis of using RLS in Neural Fuzzy Systems

Analysis of using RLS in Neural Fuzzy Systems

引用

IEEE International Conference on Systems, Man and Cybernetics (SMC)

作者： Yeh, Jen-Wei Su, Shun-Feng Rudas, Imre Natl Taiwan Univ Sci & Technol Dept Elect Engn Taipei Taiwan Obuda Univ Budapest Hungary

ISBN: (纸本)9781457706530

In this study, we continue our analysis on the use of RLS in neural fuzzy systems. The recursive least square (RLS) algorithms can have great learning performance for neural fuzzy networks. From our previous work, it can be observed that the advantages of using RLS instead of using BP are not so obvious. For the use of forgetting factor in RLS, the idea is to account for the effects of the change in the premise part. In this study, we have observed that the use of a forgetting factor can still have some advantages when the premise part is fixed. The idea is similar to the used of Widrow-Hoff learning concept in the backpropagation learning algorithm. From our experiments, a strong forgetting factor (smaller value) can let the consequent part trace the error in the learning phase. But the testing error becomes very large. When the system capacity is sufficient, a forgetting factor will improve both in the learning phase and in the testing phase. Finally, the initial value of the covariance matrix is considered. The learning capacity will rise when the initial value increases. But it will increase the error tracing phenomenon in the consequent part too. But it is opposite in a system with less learning capacity.

关键词： Covariance matrix Estimation Fuzzy systems Neural networks Numerical models RLS Testing Training Widrow-Hoff learning concept backpropagation backpropagation learning algorithm covariance matrices covariance matrix error detection error tracing phenomenon

来源：评论

学校读者我要写书评

暂无评论

Neural network modelling and sliding mode control of a hydrocarbon degradation process

Neural network modelling and sliding mode control of a hydro...

引用

15th IASTED International Conference on Applied Simulation and Modelling

作者： Baruch, Ieroham S. Mariaca G., Carlos R. Cruz, Israel Barrera C., Josefina IPN CINVESTAV Dept Automat Control Ave IPN 2508 AP 14-470 Mexico City 07360 DF Mexico IPN CINVESTAV Dept Biotechnol & Bioengn Mexico City 07360 DF Mexico

ISBN: (纸本)0889865590

This paper proposes the use of a Recurrent Neural Network (RNN) for modeling a hydrocarbon degradation process carried out in a biopile system. The proposed RNN model has seven inputs, five outputs and twelve neurons in the hidden layer, with global and local feedbacks. The learning algorithm is a modified version of the dynamic backpropagation. The obtained RNN model is used to design a sliding mode control. The simulation results obtained with the RNN model learning and control exhibit a good convergence and tracking.

关键词： recurrent neural networks model backpropagation learning algorithm Kalman filter sliding mode control hydrocarbon degradation and biopile system

来源：评论

学校读者我要写书评

暂无评论

A Fast learning algorithm with Promising Convergence Capability

A Fast Learning Algorithm with Promising Convergence Capabil...

引用

International Joint Conference on Neural Networks (IJCNN)

作者： Cheung, Chi-Chung Ng, Sin-Chun Lui, Andrew K. Xu, Sean Shensheng Hong Kong Polytech Univ Dept Elect & Informat Engn Hong Kong Hong Kong Peoples R China Open Univ Sch Sci Technol Hong Kong Peoples R China

ISBN: (纸本)9781424496365

backpropagation (BP) learning algorithm is the most widely supervised learning technique which is extensively applied in the training of multi-layer feed-forward neural networks. Many modifications of BP have been proposed to speed up the learning of the original BP. However, these modifications sometimes cannot converge properly due to the local minimum problem. This paper proposes a new algorithm, which provides a systematic approach to make use of the characteristics of different fast learning algorithms so that the convergence of a learning process is promising with a fast learning rate. Our performance investigation shows that the proposed algorithm always converges with a fast learning rate in two popular complicated applications whereas other popular fast learning algorithms give very poor global convergence capabilities in these two applications.

关键词： Breast cancer Convergence Iris Iris recognition Neural networks Neurons Switches backpropagation backpropagation learning algorithm fast learning algorithm feedforward neural nets learning technique local minimum problem multilayer feed-forward neural networks promising convergence capability

来源：评论

学校读者我要写书评

暂无评论

Application of machine learning techniques for solving real world business problems: the case study - target marketing of insurance policies

Application of machine learning techniques for solving real ...

引用

作者： Juozenaite, Ineta Universidade Nova de Lisboa

学位级别：硕士

The concept of machine learning has been around for decades, but now it is becoming more and more popular not only in the business, but everywhere else as well. It is because of increased amount of data, cheaper data storage, more powerful and affordable computational processing. The complexity of business environment leads companies to use data-driven decision making to work more efficiently. The most common machine learning methods, like Logistic Regression, Decision Tree, Artificial Neural Network and Support Vector Machine, with their applications are reviewed in this work. Insurance industry has one of the most competitive business environment and as a result, the use of machine learning techniques is growing in this industry. In this work, above mentioned machine learning methods are used to build predictive model for target marketing campaign of caravan insurance policies to achieve greater profitability. Information Gain and Chi-squared metrics, Regression Stepwise, R package "Boruta", Spearman correlation analysis, distribution graphs by target variable, as well as basic statistics of all variables are used for feature selection. To solve this real-world business problem, the best final chosen predictive model is Multilayer Perceptron with backpropagation learning algorithm with 1 hidden layer and 12 hidden neurons.

关键词： Machine learning Logistic Regression Decision Tree CART Artificial Neural Network Multilayer Perceptron backpropagation learning algorithm Support Vector Machine Kernel Gaussian Radial Basis Function Target Marketing Insurance Policies Information Gain Chi-squared metric Regression Stepwise Stratified Data Partitioning Spearman Correlation

来源：评论

学校读者我要写书评

暂无评论

Automatic digital modulation recognition using artificial neural networks

Automatic digital modulation recognition using artificial ne...

引用

International Conference on Neural Networks and Signal Processing

作者： Zhao, YQ Ren, GH Wang, XX Wu, ZL Gu, XM Harbin Inst Technol Dept Elect & Commun Engn Harbin 150001 Heilongjiang Peoples R China

ISBN: (纸本)0780377028

This paper presents a modified structure and learning algorithm of artificial neural networks (ANN) for recognizing baseband signal modulation types in the presence of additive white Gaussian noise. The new method employs a layer with less output nodes and an error back propagation learning algorithm with momentum to improve the recognition performance. Simulation results and performance evaluation of the ANN are given and it is shown that the benefits of the developed method are that its structure is simple and it performs well at low signal to noise ratio (SNR) with high overall success rates.

关键词： modulation neural nets AWGN backpropagation computational complexity signal processing automatic digital modulation recognition artificial neural networks baseband signal modulation additive white Gaussian noise backpropagation learning algorithm signal to noise ratio computational complexity

来源：评论

学校读者我要写书评

暂无评论

Time Series Forecasting With Orthogonal Endocrine Neural Network Based on Postsynaptic Potentials

引用

JOURNAL OF DYNAMIC SYSTEMS MEASUREMENT AND CONTROL-TRANSACTIONS OF THE ASME 2017年第4期139卷 041006-041006页

作者： Milovanovic, Miroslav Antic, Dragan Milojkovic, Marko Nikolic, Sasa S. Spasic, Miodrag Peric, Stanisa Univ Nis Fac Elect Engn Dept Control Syst Aleksandra Medvedeva 14 Nish 18000 Serbia

This paper presents a new type of endocrine neural network (ENN). ENN utilizes artificial glands which enable the network to be adaptive to external disturbances. Sensitivity is controlled by the hormone decay rate and the value of the sensitivity parameter. The network presented in this paper is improved by making the sensitivity parameter selftuning and implementing orthogonal activation functions inside the network structure. Automatic tuning is performed on the basis of the biological principle of postsynaptic potentials by implementing inhibitory and excitatory glands inside the standard backpropagation learning algorithm of developed orthogonal ENN. These additional network functionalities enable extra sensitivity to external conditions and an additional network feature of activation sharpening. The network was tested on real-time series of experimental data with a purpose to forecast exchange rate of the three widely used international currencies.

关键词： endocrine neural network time series forecasting orthogonal functions postsynaptic potential sensitivity parameter backpropagation learning algorithm

来源：评论

学校读者我要写书评

暂无评论

Applications of Ground-Penetrating Radar (GPR) to Detect Hidden Beam Positions

引用

JOURNAL OF TESTING AND EVALUATION 2017年第3期45卷 911-921页

作者： Kilic, Gokhan Izmir Univ Econ Dept Civil Engn Izmir Turkey

Ground-penetrating radar (GPR) uses electromagnetic waves to investigate the structures. In this investigation method, an electromagnetic wave is transmitted using an antenna and the received signal is recorded. Detection of beam positions in this GPR data requires the skills of a trained human operator. This study utilized a multi-layer neural network to detect beam positions in the GPR data. The visual description and definition of GPR data has major disadvantages and a neural network has been studied to overcome these shortcomings. A set of 32,740 training vectors with a length of 64 data was implemented to train the neural network. A new set of 16,370 testing vectors with a length of 64 data was then prepared to test the performance. Testing results suggest that the neural network is promising methods for the detection of beam positions in the GPR data.

关键词： backpropagation learning algorithm Bayes optimal decision rule Gram-Charlier series GPR and data processing neural network

来源：评论

学校读者我要写书评

暂无评论

Distributed Neural Observer-Based Formation Strategy of Non-Affine Nonlinear Multi-Agent Systems with Unknown Dynamics

引用

JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS 2021年第5期30卷

作者： Manouchehri, Pejman Ghasemi, Reza Toloei, Alireza Mohammadi, Fazel Islamic Azad Univ Elect Engn Dept Damavand Branch Tehran Iran Univ Qom Elect Engn Dept Qom Iran Shahid Beheshti Univ Fac Engn Mech Engn Dept Tehran Iran Univ Windsor Elect & Comp Engn Dept Windsor ON N9B 1K3 Canada

The state estimation in Multi-Agent Systems (MASs) is a challenging problem. This is due to the fact that (1) controlling nonaffine nonlinear MASs is a difficult task and also (2) the agents in MASs have direct impacts on each other. This paper presents a new distributed Neural Networks (NN) observer for the nonlinear dynamical model of MASs with nonaffine unknown dynamical agents. The proposed scheme uses the backpropagation learning algorithm to estimate the unknown nonlinear functions of the agents. Compared with the previous studies, which primarily concentrated on the observer design for Multiple Input Multiple Output (MIMO) systems, the proposed method is applied to nonaffine nonlinear MASs. The advantages of this method are the overall stability, the fast convergence of the observer error to zero and the robustness against both uncertainties and disturbances. Nonlinear flexible-joint robots and nonlinear dynamic duffing chaotic systems are simulated to demonstrate the effectiveness and robustness of the proposed method. The proposed method is also compared with the Luenberger observer. The guaranteed stability, better performance in the presence of agents' uncertainties, robustness against disturbances are the main advantages of the proposed method compared with the traditional observer.

关键词： backpropagation learning algorithm chaotic system distributed neural observer nonaffine nonlinear systems Multi-Agent System (MAS)

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：