空间搜索优化多项式神经网络分类器：结合策略调整与数据预处理

9 浏览量更新于2024-08-26 收藏 527KB PDF 举报

"这篇论文提出了一种优化的多项式神经网络分类器，通过结合空间搜索优化策略和数据预处理技术来提升分类性能。该方法针对神经网络分类器在设计时通常面临的三个挑战：判别函数的选择、大量参数的设定以及高维训练数据的处理。论文作者是Wei Huang和Sung-Kwun Oh，发表在JElectrEngTechnol期刊2017年的第12卷第2期，911-917页。" 在神经网络分类器的设计中，有三个关键方面需要考虑。首先，判别函数的选择对于分类效果至关重要，它定义了不同类别的边界。其次，设计过程中涉及大量的参数，如神经元数量、连接权重、学习速率等，这些参数的设置直接影响到模型的性能。最后，高维度的训练数据常常带来计算复杂性和过拟合的问题。论文提出的优化多项式神经网络分类器（PNNC）结合了空间搜索优化策略，这种策略能够探索参数空间，寻找最优的网络结构和参数设置。空间搜索通过算法自动调整网络的复杂性，以适应数据集的特性，从而提高分类准确率。此外，PNNC还利用了数据预处理技术，如主成分分析（PCA）或线性判别分析（LDA），来降低数据的维度，减少冗余特征，提高模型的训练效率和泛化能力。 PCA是一种统计方法，能将高维数据转换为一组线性不相关的低维向量，保留原始数据的主要信息。LDA则是一种有监督的降维技术，旨在找到能最大化类别间距离同时最小化类别内距离的投影方向。这些预处理技术的应用有助于简化模型的训练过程，减轻过拟合风险，并可能提升对噪声数据的鲁棒性。同时，论文中的“同步调优策略”（simultaneous tuning strategy）意味着在进行空间搜索优化时，不仅优化模型的结构，也同时调整数据预处理的参数，以达到整体性能的最佳平衡。这种方法确保了模型设计和数据处理两个阶段的协同优化，提高了分类器的整体效能。这项工作为神经网络分类器的设计提供了一个新的视角，通过整合优化技术、数据预处理和智能搜索策略，有效地解决了高维数据和大量参数问题，提升了多项式神经网络的分类性能。这种方法对于处理复杂和高维数据的机器学习任务具有潜在的应用价值。

J Electr Eng Technol.2017; 12(2): 911-917

https://doi.org/10.5370/JEET.2017.12.2.911

911

This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/

licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Optimized Polynomial Neural Network Classifier Designed with the Aid

of Space Search Simultaneous Tuning Strategy and Data Preprocessing

Techniques

Wei Huang* and Sung-Kwun Oh

†

Abstract – There are generally three folds when developing neural network classifiers. They are as

follows: 1) discriminant function; 2) lots of parameters in the design of classifier; and 3) high

dimensional training data. Along with this viewpoint, we propose space search optimized polynomial

neural network classifier (PNNC) with the aid of data preprocessing technique and simultaneous

tuning strategy, which is a balance optimization strategy used in the design of PNNC when running

space search optimization. Unlike the conventional probabilistic neural network classifier, the

proposed neural network classifier adopts two type of polynomials for developing discriminant functions.

The overall optimization of PNNC is realized with the aid of so-called structure optimization and

parameter optimization with the use of simultaneous tuning strategy. Space search optimization

algorithm is considered as a optimize vehicle to help the implement both structure and parameter

optimization in the construction of PNNC. Furthermore, principal component analysis and linear

discriminate analysis are selected as the data preprocessing techniques for PNNC. Experimental results

show that the proposed neural network classifier obtains better performance in comparison with some

other well-known classifiers in terms of accuracy classification rate.

Keywords: Polynomial Neural Network Classifier (PNNC), Principal component analysis, Linear

discriminate analysis, Space search optimization, Simultaneous tuning strategy, Data preprocessing

technique

1. Introduction

In the past decades, lot of neural network classifier have

been developed in many research fields. Among these

neural network classifiers, lot of architectures, learning

abilities, and applications have been well documented, c.f.

[4, 5], and [6-8]. Generally, there are three aspects in the

design of neural network classifiers. They are as follows:

Discriminate function;

Lots of parameters in the design of classifier;

High dimensional training data.

Linear discriminate function is generally used when

constructing the neural network classifiers. However, such

linear discriminate functions have a relatively simple

geometry. For example, in the design radial basis function

neural network classifier, such simple geometry is implied

by the limited geometric variability of RBFs forming the

hidden layer of the radial basis function neural network.

Some enhancements have been shown by using two order

polynomial discriminate functions [4], the problem of

using high order discriminate function remains open.

In the design of neural network classifiers, there are

generally several parameters needed to be estimated. A

variety of evolutionary optimization algorithms such as

genetic algorithm [9], particle swarm optimizer [10],

differential evolution [11-12] have been successfully applied

to adjust parameters when developing the neural network

classifiers. For example, in the design of RBF neural

network classifier, the parameters including the number of

RBFs, the coordinates of RBF centers, the widths of RBFs

are optimized by using evolutionary algorithms. In spite

of advantages, these evolutionary algorithms cannot free

from limitations. The first apparent limitation is that these

evolutionary algorithms search solution space based on

pure random mechanism. The second apparent limitation

is that a problem of how to balance the structure

optimization and parameter optimization, which are

commonly encountered in the optimization of classifiers.

Such problem plays an important role in the performance

of neural network classifiers. Nevertheless, most of

existing classifiers have not consider the above two

limitations.

In this study, we propose a novel polynomial neural

network classifier (PNNC) based on space search optimi-

zation algorithm using simultaneous tuning strategy to

overcome all aforementioned limitations. The proposed

† Corresponding Author: Dep. of Electrical Engineering, The

University of Suwon, Korea. (ohsk@suwon.ac.kr)

* School of Computer and Communication Engineering, Tianjin

University of Technology, China. / State Key Laboratory of Digital

Manufacturing Equipment and Technology, Huazhong University of

Science and Technology, China. (huangwabc@163.com)

Received: July 11, 2016; Accepted: October 31, 2016

ISSN(Print) 1975-0102

ISSN(Online) 2093-7423

下载后可阅读完整内容，剩余6页未读，立即下载

weixin_38692100

粉丝: 3
资源: 870

空间搜索优化多项式神经网络分类器：结合策略调整与数据预处理

基于数据预处理和空间搜索优化的多项式神经网络分类器

SVM神经网络中的参数优化---提升分类器性能(matlab实现).zip

数据预处理与空间搜索优化的多项式神经网络分类器

人工智能-项目实践-数据预处理-“中国工业大数据创新竞赛” 数据预处理与分类

MATLAB神经网络和优化算法：29 SVM神经网络中参数优化提升分类器性能.zip

SVM参数优化技术：提升神经网络分类器性能

MATLAB神经网络参数优化技巧：提升SVM分类器性能

MATLAB中SVM神经网络参数优化提升分类器性能案例分析

数据预处理与十大经典算法解析

【数据预处理黄金法则】：为Python神经网络算法打造高效数据集

最新资源