检索结果-内蒙古大学图书馆

Feature subset selection by gravitational search algorithm optimization

INFORMATION SCIENCES 2014年 281卷 128-146页

作者： Han, XiaoHong Chang, XiaoMing Quan, Long Xiong, XiaoYan Li, JingXia Zhang, ZhaoXia Liu, Yi Taiyuan Univ Technol Key Lab Adv Transducers & Intelligent Control Sys Minist Educ China Taiyuan Shanxi Peoples R China

A new method for feature subset selection in machine learning, FSS-MGSA (Feature Subset Selection by Modified Gravitational Search algorithm), is presented. FSS-MGSA is an evolutionary, stochastic search algorithm based on the law of gravity and mass interactions, and it can be executed when domain knowledge is not available. A wrapper approach, over Naive-Bayes, ID3, K-Nearest Neighbor and Support Vector Machine learning algorithms, is used to evaluate the goodness of each visited solution. The key to the success of the MGSA is to utilize the piecewise linear chaotic map for increasing its diversity of species, and to use sequential quadratic programming for accelerating local exploitation. Promising results are achieved in a variety of tasks where domain knowledge is not available. The experimental results show that the proposed method has the ability of selecting the discriminating input features correctly and can achieve high accuracy of classification, which is comparable to or better than well-known similar classifier systems. Furthermore, the MGSA is tested on ten functions provided by CEC 2005 special session and compared with various modified Gravitational Search algorithm, Particle Swarm Optimization, and Genetic algorithm. The obtained results confirm the high performance of the MGSA in solving various problems in optimization. (C) 2014 Elsevier Inc. All rights reserved.

关键词： Feature subset selection Classification Gravitational search algorithm learning algorithm Sequential quadratic programming Chaotic map

来源：评论

学校读者我要写书评

暂无评论

A Type of Directly Self-adaptive PID Controller Based on Neural Network

A Type of Directly Self-adaptive PID Controller Based on Neu...

引用

2015 4th National Conference on Electrical,Electronics and Computer Engineering(NCEECE 2015)

作者： Han Chong Commercial Aircraft Corporation of China

ISBN: (纸本)9781510819085

A type of directly self-tuning PID controller based on neural network is proposed in this paper. Its main characteristic is that it no longer includes independent PID controller and put neural network and the law of PID controller together. Showing the learning algorithm of this neural network controller and analyzing the stability of this control system. The simulated results prove that this kind of control system is more adaptive and robust.

关键词： neural network PID control nonlinearity learning algorithm

来源：评论

学校读者我要写书评

暂无评论

DIFFUSION learning algorithmS FOR FEEDFORWARD NEURAL NETWORKS

引用

CYBERNETICS AND SYSTEMS ANALYSIS 2013年第3期49卷 334-346页

作者： Skorohod, B. A. Sevastopol Natl Tech Univ Sevastopol Ukraine

The problem of training feedforward neural networks is considered. To solve it, new algorithms are proposed. They are based on the asymptotic analysis of the extended Kalman filter (EKF) and on a separable network structure. Linear weights are interpreted as diffusion random variables with zero expectation and a covariance matrix proportional to an arbitrarily large parameter lambda. Asymptotic expressions for the EKF are derived as lambda -> infinity. They are called diffusion learning algorithms (DLAs). It is shown that they are robust with respect to the accumulation of rounding errors in contrast to their prototype EKF with a large but finite lambda and that, under certain simplifying assumptions, an extreme learning machine (ELM) algorithm can be obtained from a DLA. A numerical example shows that the accuracy of a DLA may be higher than that of an ELM algorithm.

关键词： feedforward neural network learning algorithm extended Kalman filter

来源：评论

学校读者我要写书评

暂无评论

Tunable and Generic Problem Instance Generation for Multi-objective Reinforcement learning

Tunable and Generic Problem Instance Generation for Multi-ob...

引用

IEEE Symposium on Adaptive Dynamic Programming and Reinforcement learning (ADPRL)

作者： Garrett, Deon Bieger, Jordi Throisson, Kristinn R. Reykjavik Univ Iceland Inst Intelligent Machines Reykjavik Iceland Reykjavik Univ Reykjavik Iceland

ISBN: (纸本)9781479945528

A significant problem facing researchers in reinforcement learning, and particularly in multi-objective learning, is the dearth of good benchmarks. In this paper, we present a method and software tool enabling the creation of random problem instances, including multi-objective learning problems, with specific structural properties. This tool, called Merlin (for Multi-objective Environments for Reinforcement learning), provides the ability to control these features in predictable ways, thus allowing researchers to begin to build a more detailed understanding about what features of a problem interact with a given learning algorithm to improve or degrade the algorithm's performance. We present this method and tool, and briefly discuss the controls provided by the generator, its supported options, and their implications on the generated benchmark instances.

关键词： learning (artificial intelligence) software tools Merlin learning algorithm multiobjective environments for reinforcement learning multiobjective learning problem multiobjective reinforcement learning problem facing researcher random problem instance software tool structural property Benchmark testing Correlation Covariance matrices Generators Heuristic algorithms learning (artificial intelligence) Optimization Software Tools Neurofibromin 2 Benchmark testing Structural properties dynamos learning algorithms learning (artificial intelligence) Heuristic algorithms variance covariance matrix Covariance matrix

来源：评论

学校读者我要写书评

暂无评论

A learning algorithm considering derivative information into neural networks

引用

International Review on Computers and Software 2010年第3期5卷 324-330页

作者： Wu, Shuang Li, Haibin Mechanics Department of Inner Mongolia University of Technology China

In this paper, a learning algorithm considering derivative information is proposed for neural networks. Based on backpropagation (BP) algorithm, this algorithm takes the derivative information of the samples into account. Because the partial derivative of the output variable of neural networks to the input variable can be obtained from the weights of the networks, the error between the calculated value and the theoretical value of the partial derivative can be reduced by adjusting the weights of the networks. This algorithm can effectively avoid the over-fitting and improve the generalization capability for the networks. The neural networks considering derivative information has been shown by experimental results to have better generalization performance than BP networks. © 2010 Praise Worthy Prize S.r.l. - All rights reserved.

关键词： Derivative information Generalization capability learning algorithm Neural networks

来源：评论

学校读者我要写书评

暂无评论

PREVIEW ON STRUCTURES AND algorithmS OF DEEP learning 11

PREVIEW ON STRUCTURES AND ALGORITHMS OF DEEP LEARNING

引用

11th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)

作者： Kuang, Ping Cao, Wei-Na Wu, Qiao Univ Elect Sci & Technol China Sch Comp Sci & Engn Chengdu 611731 Peoples R China

ISBN: (纸本)9781479972081

Deep learning proposed by Hinton et al is a new learning algorithm of multi-layer neural network, and it is also a new study field in machine learning. This paper describes the structures and advantages to shallow learning of deep learning, and analyzes current popular learning algorithm in detail. Finally, this paper analyzes research directions and future prospects of deep learning.

关键词： Deep learning learning algorithm machine learning

来源：评论

学校读者我要写书评

暂无评论

Performance analysis of statistical optimal data fusion algorithms

引用

INFORMATION SCIENCES 2014年 277卷 808-824页

作者： Xia, Youshen Leung, Henry Fuzhou Univ Coll Math & Comp Sci Fuzhou 350002 Peoples R China Univ Calgary Dept Elect & Comp Engn Calgary AB T2N 1N4 Canada

Data fusion is widely used in biometric, multi-media signal and image processing, and wireless sensor networks. Optimal fusion techniques are developed to perform fusion under noisy environments. However, the statistical analysis carried out to evaluate the relative merits of these optimal methods is very less. The aim of this paper is to fill this gap by evaluating four statistical optimal data fusion methods namely, the linearly constrained least squares (LCLS) fusion method, the covariance intersection (Cl) fusion method, the linearly constrained least absolute deviation (CLAD) fusion method, and the constrained least square (CLS) fusion method. The CLS fusion method presented here is an improved version of the CLAD fusion method. We further analyze the performances of these four methods in terms of optimality, unbiased estimation, robustness, and complexity. Simulations are used to validate the performance of these fusion algorithms. (C) 2014 Elsevier Inc. All rights reserved.

关键词： Optimal data fusion Signal and image fusion learning algorithm Performance analysis

来源：评论

学校读者我要写书评

暂无评论

Artificial neural network approach to the fuzzy Abel integral equation problem

引用

JOURNAL OF INTELLIGENT & FUZZY SYSTEMS 2014年第1期27卷 83-91页

作者： Jafarian, Ahmad Nia, Safa Measoomy Islamic Azad Univ Dept Math Urmia Branch Orumiyeh Iran

Artificial neural networks (ANNs) have a large appeal to many researchers due to their feature to simulate and solve different kinds of problems that do not have algorithmic solutions. The outlined in this paper is an efficient and robust collocation method based on the ANNs and Bernstein polynomials intended for the fuzzy Abel integral equation problem. To do this, first truncated Bernstein-series polynomial of the solution function is substituted in parametric form of the given fuzzy problem. Then an architecture of ANNs namely the feed-back neural nets is designed to determine values for the unknown coefficients. Eventually, the proposed method is implemented on some numerical examples, and also is compared with an usual and classical technique.

关键词： Fuzzy Abel integral equation bivariate Bernstein polynomial feed-back neural network cost function learning algorithm

来源：评论

学校读者我要写书评

暂无评论

A Handoff algorithm Based on Estimated Load for Dense Green 5G Networks

A Handoff Algorithm Based on Estimated Load for Dense Green ...

引用

IEEE Global Communications Conference

作者： Atefeh Hajijamali Arani Mohammad Javad Omidi Abolfazl Mehbodniya Fumiyuki Adachi Dept. of Electrical and Computer Engineering Isfahan University of Technology Dept. of Communication Engineering Graduate School of Engineering Tohoku University

ISBN: (纸本)9781479959532

Two important challenges facing 5G are energy efficiency and mobile users' mobility in heterogeneous wireless networks (HetNets). One of the important techniques for improving energy efficiency is base station (BS)'s switching between ON and OFF modes which allows the BS to turn off some its components in lower load situations. In this paper, we address user's seamless mobility problem and propose a handoff (HO) algorithm based on BS's estimated load. The proposed HO algorithm based on estimated load (PHA-EL) balances load by imposing HOs from highly loaded BSs to lightly loaded BSs. When a BS is overloaded, the user's quality of service (QoS) will degrade and therefore the PHA-EL is used to improve system throughput. The PHA-EL algorithm is combined with BSs which are able to switch between ON and OFF modes (PHA-EL/ON-OFF switching) in order to improve the energy efficiency of the system. Therefore, this algorithm achieves both energy- and spectral- efficiency. Simulation results indicate that the proposed algorithm yields better performance in terms of average number of HOs, average load per BS and average payoff per BS, compared to baseline algorithm.

关键词： Energy efficiency Heterogeneous wireless network Handoff learning algorithm Energy efficiency handoff heterogeneous wireless network learning algorithms Bahamas LOAD FBXW11 gene

来源：评论

学校读者我要写书评

暂无评论

Generalization ability of fractional polynomial models

引用

NEURAL NETWORKS 2014年 49卷 59-73页

作者： Lei, Yunwen Ding, Lixin Ding, Yiming Wuhan Univ Sch Comp State Key Lab Software Engn Wuhan 430072 Peoples R China Chinese Acad Sci Wuhan Inst Phys & Math Wuhan 430071 Peoples R China

In this paper, the problem of learning the functional dependency between input and output variables from scattered data using fractional polynomial models (FPM) is investigated. The estimation error bounds are obtained by calculating the pseudo-dimension of FPM, which is shown to be equal to that of sparse polynomial models (SPM). A linear decay of the approximation error is obtained for a class of target functions which are dense in the space of continuous functions. We derive a structural risk analogous to the Schwartz Criterion and demonstrate theoretically that the model minimizing this structural risk can achieve a favorable balance between estimation and approximation errors. An empirical model selection comparison is also performed to justify the usage of this structural risk in selecting the optimal complexity index from the data. We show that the construction of FPM can be efficiently addressed by the variable projection method. Furthermore, our empirical study implies that FPM could attain better generalization performance when compared with SPM and cubic splines. (C) 2013 Elsevier Ltd. All rights reserved.

关键词： learning algorithm learning theory Fractional polynomial Model selection Approximation theory

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：