检索结果-内蒙古大学图书馆

training algorithms for Radial Basis Function Networks to tackle learning processes with imbalanced data-sets

APPLIED SOFT COMPUTING 2014年第1期25卷 26-39页

作者： Perez-Godoy, M. D. Rivera, Antonio J. Carmona, C. J. del Jesus, M. J. Univ Jaen Dept Comp Sci Jaen 23071 Spain

Nowadays, many real applications comprise data-sets where the distribution of the classes is significantly different. These data-sets are commonly known as imbalanced data-sets. Traditional classifiers are not able to deal with these kinds of data-sets because they tend to classify only majority classes, obtaining poor results for minority classes. The approaches that have been proposed to address this problem can be categorized into three types: resampling methods, algorithmic adaptations and cost sensitive techniques. Radial Basis Function Networks (RBFNs), artificial neural networks composed of local models or RBFs, have demonstrated their efficiency in different machine learning areas. Centers, widths and output weights for the RBFs must be determined when designing RBFNs. Taking into account the locally tuned response of RBFs, the objective of this paper is to study the influence of global and local paradigms on the weights training phase, within the RBFNs design methodology, for imbalanced data-sets. Least Mean Square and the Singular Value Decomposition have been chosen as representatives of local and global weights training paradigms respectively. These learning algorithms are inserted into classical RBFN design methods that are run on imbalanced data-sets and also on these data-sets preprocessed with re-balance techniques. After applying statistical tests to the results obtained, some guidelines about the RBFN design methodology for imbalanced data-sets are provided. (C) 2014 Elsevier B.V. All rights reserved.

关键词： Imbalanced data sets Radial Basis Function Networks training algorithms Evolutionary computation

来源：评论

学校读者我要写书评

暂无评论

training algorithms FOR BACKPROPAGATION NEURAL NETWORKS WITH OPTIMAL DESCENT FACTOR

引用

ELECTRONICS LETTERS 1990年第20期26卷 1698-1700页

作者： YU, XH CHENG, SX Dept. of Radio Eng. Southeast Univ. Nanjing China

Poor convergence of existing training algorithms prevents wide applications of backpropagation neural networks. Several new training algorithms with very fast convergence are presented. They all use derivative information to efficiently estimate the optimal descent factors, thus providing the fastest descent on the mean squared error in the descent directions that characterise the algorithms. Simulation results are illustrated.

关键词： backpropagation neural networks computerised signal processing derivative information fast convergence mean squared error network analysis neural nets optimal descent factor optimal descent factors signal processing simulation results subroutines training algorithms

来源：评论

学校读者我要写书评

暂无评论

Combining various training and adaptation algorithms for ensemble few-shot classification

引用

NEURAL NETWORKS 2025年 185卷 107211页

作者： Jiang, Zhen Tang, Na Sun, Jianlong Zhan, Yongzhao Jiangsu Univ Sch Comp Sci & Commun Engn Zhenjiang Peoples R China

To mitigate the shortage of labeled data, Few-Shot Classification (FSC) methods train deep neural networks (DNNs) on a base dataset with sufficient labeled data, and then adapt them to target tasks using a few labeled data. Despite notable progress, a single FSC model remains prone to high variance and low confidence. As a result, ensemble FSC has garnered increasing attention. However, the limited labeled data and the high computational cost associated with DNNs present significant challenges for ensemble FSC methods. This paper presents a novel ensemble method that generates multiple FSC models via combining various training and adaptation algorithms. Due to the reuse of training phases, the proposed method significantly reduces the learning cost while generating base models with greater diversity. To further minimize reliance on labeled data, we provide each model with pseudo-labeled data selected by the majority vote of other models. Compared with self-training style methods, this "one-vs-others" learning strategy effectively reduces pseudo-label noise and confirmation bias. Finally, we conduct extensive experiments on miniImageNet, tieredImageNet and CUB datasets. The experimental results demonstrate that our method outperforms other state-of-the-art FSC methods. Especially, our method achieves the greatest improvement in the performance of base models. The source code and related models are available at https://***/tn1999tn/Ensemble-FSC/tree/master.

关键词： Few-shot classification Ensemble learning training algorithms Adaptation algorithms Pseudo-labeled data

来源：评论

学校读者我要写书评

暂无评论

River Suspended Sediment Prediction Using Various Multilayer Perceptron Neural Network training algorithms-A Case Study in Malaysia

引用

WATER RESOURCES MANAGEMENT 2012年第7期26卷 1879-1897页

作者： Mustafa, M. R. Rezaur, R. B. Saiedi, S. Isa, M. H. Univ Teknol Petronas Dept Civil Engn Bandar Seri Iskandar 31750 Tronohperak Dar Malaysia Golder Associates Calgary AB T2A 7W5 Canada

Estimation of suspended sediment discharge in rivers has a vital role in dealing with water resources problems and hydraulic structures. In this study, a Multilayer Perceptron (MLP) feed forward neural network with four different training algorithms was used to predict the suspended sediment discharge of a river (Pari River at Silibin) in Peninsular Malaysia. The training algorithms are Gradient Descent (GD), Gradient Descent with Momentum (GDM), Scaled Conjugate Gradient (SCG), and Levenberg Marquardt (LM). Different statistical measures, time of convergence and number of epochs to reach the required accuracy were used to evaluate the performance of training algorithms. The analysis showed that SCG and LM performed better than GD and GDM. While the performance of the superior algorithms (i.e., SCG and LM) is similar, LM required considerably shorter time of convergence. It was concluded that both training algorithms SCG and LM could be recommended for suspended sediment prediction using MLP networks. However, LM was the faster (1/7 of SCG convergence time) of the two algorithms.

关键词： Multilayer perceptron neural network training algorithms Discharge Suspended sediment Prediction Modeling

来源：评论

学校读者我要写书评

暂无评论

Linear neural network training algorithms for real-world benchmark problems

引用

INTERNATIONAL JOURNAL OF COMPUTER MATHEMATICS 2002年第11期79卷 1149-1167页

作者： Goulianas, K Adamopoulos, M Katsavounis, S Fragakis, C Tsouros, CC Technol Educ Inst Thessaloniki Dept Informat Thessaloniki Greece Univ Macedonia Dept Informat Thessaloniki Greece Aristotle Univ Thessaloniki Fac Engn Thessaloniki Greece

This paper describes the Adaptive Steepest Descent (ASD) and Optimal Fletcher-Reeves (OFR) algorithms for linear neural network training. The algorithms are applied to well-known pattern classification and function approximation problems, belonging to benchmark collection Proben1. The paper discusses the convergence behavior and performance of the ASD and OFR training algorithms by computer simulations and compares the results with those produced by linear-RPROP method.

关键词： neural nets training algorithms iterative methods

来源：评论

学校读者我要写书评

暂无评论

Neural network based prediction of mechanical properties of particulate reinforced metal matrix composites using various training algorithms

引用

MATERIALS & DESIGN 2007年第2期28卷 616-627页

作者： Koker, Rasit Altinkok, Necat Demir, Adem Sakarya Univ Fac Engn Dept Comp Engn TR-54187 Sakarya Turkey Sakarya Univ Tech Educ Fac Dept Met Educ TR-54187 Sakarya Turkey

Recently, artificial neural networks are used as an. interdisciplinary tool m many applications. There are various training algorithms used in neural network applications. In this study, it is aimed to investigate the effect of various training algorithms on learning performance of the neural networks on the prediction of bending strength and hardness behaviour of particulate reinforced Al-Si-Mg metal matrix composites (MMCs). Al2O3/SiC particulates reinforced MMC was produced by using stir casting process. Al2O3/SiC powder mix obtained from Al2O3/SiC ceramic cake, which was produced firing of aluminium sulphate aqueous solution and SiC mix before stir casting. In the experimental processes, during fabrication, stirring was applied to create a vortex for addition reinforcing particles and the production of aluminium alloy metal-matrix composites. 10 vol.% of dual ceramic powder with different SiC particle size range mix was inserted in liquid aluminium by using stir casting under Ar pressure to obtain dual particulates reinforced MMCs. This mixing was achieved successfully;microstructure, bending strength and hardness properties of the composites were tested. Bending strength and hardness behaviour were predicted with four different training algorithms using a back-propagation neural network. The training and test sets of the neural network were initially prepared using experimental results that were obtained and recorded in a file on a computer. Test results revealed that bending strength and hardness resistance of the composites increased with decrease in ductility, with decrease size of the reinforcing SiC particulates in the aluminium alloy metal matrix. In the training and test modules of the neural network, different SiC particles size (mu m) range was used as input and bending strength and hardness behaviour were used as output in the produced MMCs. After the preparation of the training set, the neural network was trained using four different training algorit

关键词： metal matrix composites production stir casting mechanical properties artificial neural networks training algorithms

来源：评论

学校读者我要写书评

暂无评论

FAST training algorithms FOR MULTILAYER NEURAL NETS

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS 1991年第3期2卷 346-354页

作者： BRENT, RP IEEE Computer Sciences Laboratory Australian National University

training a multilayer neural net by back-propagation is slow and requires arbitrary choices regarding the number of hidden units and layers. This paper describes an algorithm which is much faster than back-propagation and for which it is not necessary to specify the number of hidden units in advance. The relationship with other fast pattern recognition algorithms, such as algorithms based on k-d trees, is mentioned. The algorithm has been implemented and tested on artificial problems such as the parity problem and on real problems arising in speech recognition. Experimental results, including training times and recognition accuracy, are given. Generally, the algorithm achieves accuracy as good as or better than nets trained using back-propagation, and the training process is much faster than back-propagation. Accuracy is comparable to that of the "nearest neighbor" algorithm, which is slower and requires more storage space.

关键词： .parity pattern recognition Decision Tree training algorithms nearest neighbour Computer Sciences Perceptron NEURAL NETS multi-layer number of hidden recognition accuracy

来源：评论

学校读者我要写书评

暂无评论

training multilayered perceptrons for pattern recognition: a comparative study of four training algorithms

引用

INTERNATIONAL JOURNAL OF MACHINE TOOLS & MANUFACTURE 2001年第3期41卷 419-430页

作者： Pham, DT Sagiroglu, S Cardiff Univ Mfg Engn Ctr Intelligent Syst Lab Cardiff CF24 3TE S Glam Wales Erciyes Univ Fac Engn Dept Comp Engn Intelligent Syst Res Grp TR-38039 Kayseri Turkey

This paper presents an overview of four algorithms used for training multilayered perceptron (MLP) neural networks and the results of applying those algorithms to teach different MLPs to recognise control chart patterns and classify wood veneer defects. The algorithms studied are Backpropagation (BP), Quick-prop (QP), Delta-Bar-Delta (DBD) and Extended-Delta-Bar-Delta (EDBD). The results show that, overall, BP was the best algorithm for the two applications tested. (C) 2001 Elsevier Science Ltd. All rights reserved.

关键词： multilayered perceptrons training algorithms control chart pattern recognition wood inspection

来源：评论

学校读者我要写书评

暂无评论

Develop 24 dissimilar ANNs by suitable architectures & training algorithms via sensitivity analysis to better statistical presentation: Measure MSEs between targets & ANN for Fe-CuO/Eg-Water nanofluid

引用

PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS 2019年 519卷 159-168页

作者： Bahrami, Mehrdad Akbari, Mohammad Bagherzadeh, Seyed Amin Karimipour, Arash Afrand, Masoud Goodarzi, Marjan Islamic Azad Univ Najafabad Branch Dept Mech Engn Najafabad Iran Ton Duc Thang Univ Fac Environm & Labour Safety Sustainable Management Nat Resources & Environm R Ho Chi Minh City Vietnam

The artificial neural network optimization method is evaluated according to the experimental results of the hybrid non-Newtonian nanofluid of iron and copper oxide in a binary mixture of water and ethylene glycol concerned the mixture dynamic viscosity versus shear rate at different amounts of nanoparticles concentration and temperate. Present work novelty is demonstrated by providing 24 dissimilar ANN methods to introduce the suitable architectures and training algorithms for them. The mean squared errors (MSEs) between the targets and ANN outputs are evaluated to present the best optimization approach among them. Meanwhile the results would be supported by the appropriate sensitivity analysis to have better statistical visual presentation. (C) 2019 Elsevier B.V. All rights reserved.

关键词： ANN MSE training algorithms Non-Newtonian nanofluid Sensitivity analysis

来源：评论

学校读者我要写书评

暂无评论

Evaluation of MLP-ANN training algorithms for Modeling Soil Pore-Water Pressure Responses to Rainfall

引用

JOURNAL OF HYDROLOGIC ENGINEERING 2013年第1期18卷 50-57页

作者： Mustafa, M. R. Rezaur, R. B. Saiedi, S. Rahardjo, H. Isa, M. H. Univ Teknol Petronas Dept Civil Engn Tronoh 31750 Perak Darul Rid Malaysia Golder Associates Calgary AB T2A 7W5 Canada Nanyang Technol Univ Sch Civil & Environm Engn Singapore 639798 Singapore

Knowledge of pore-water pressure responses to rainfall is vital in slope failure and slope hydrological studies. The performance of four artificial neural network (ANN) training algorithms was evaluated to identify the training algorithm appropriate for modeling the dynamics of soil pore-water pressure responses to rainfall patterns using multilayer perceptron (MLP) ANN. The ANN model comprised eight neurons in the input layer, four neurons in the hidden layer, and a single neuron in the output layer representing an 8-4-1 ANN architecture. The training algorithms evaluated include the gradient descent, gradient descent with momentum, scaled conjugate gradient, and Levenberg-Marquardt (LM). The performance of the training algorithms was evaluated using standard performance evaluation measures-root mean square error, coefficient of efficiency, and the time and number of epochs required to reach a predefined accuracy. It was found that all the training algorithms could be used in the prediction of pore-water pressures. However, the LM algorithm required the least time and epochs for training the network and gave the minimum error during both training and testing. The LM training algorithm is therefore proposed as an ideal and fast training algorithm for modeling the dynamics of soil pore-water pressure changes in response to rainfall patterns. DOI: 10.1061/(ASCE)HE.1943-5584.0000599. (C) 2013 American Society of Civil Engineers.

关键词： Artificial neural network (ANN) training algorithms Evaluation Pore-water pressure Prediction Modeling

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：