神经网络分类器的设计与实现
毕业论文关键字：人工神经网络；BP算法；分类器；数据分析；成绩预测
ABSTRACT Currently, the classification of data due to the amount of data more and more, arebecoming a huge workload, to check if the use of human wastes a lot of resources, so manyclassification also came into being. The neural network classifier is a goodanthropomorphic computer classification, this paper design and implement some simpleneural network classifier, and introduces the basic principles of neural network classifiers,to predict student achievement.In many areas it will have a good application, for example, it can be the license platerecognition, face recognition and so on. Wide range of applications in the country, thedirection of military security, civil and economic fields, home entertainment, fieldidentification, etc. have great research value. The graduation project is intended to design anetwork before using feedforward neural excellent performance of neural networkclassifiers, and with Matlab software programming, to analyze the impact of four years ofacademic performance between students in university, and four experimental design: withdifferent hidden layer network performance prediction; 24 were compulsory performancepredicted by graduation project; all with a university English scores on the LanguageAbility to predict; circuit, the high number, such as dragging electrical performance ofautomatic control theory to predict results ; hoping to complete a can accurately predictacademic performance neural network classifier.
Keywords ： Artificial neural network; BP algorithm; Classifier; Data analysis;performance prediction
目录
第一章 绪论. 1
1.1 人工神经网络的研究目的和意义.. 1
1.2 神经网络的发展与研究现状..1
1.2.1 神经网络的发展 1
1.2.2 神经网络的研究现状 ..2
1.2.3 发展趋势及研究热点...2
1.3 BP 神经网络的优缺点介绍 ...3
1.4 本文主要工作内容....5
第二章 BP神经网络...6
2.1 BP 神经网络.... 6
2.1.1 神经网络的概述. 6
2.1.2 BP神经网络的原理 7
2.1.3 BP网络分类器的结构和机理.... 8
2.1.4 神经网络分类器的应用....9
2.2 BP算法. 14
2.3 神经网络的性质分析.. 16
第三章 MATLAB 上神经网络的实现.. 20
3.1 数据预处理.... 20
3.1.1 归一化.20
3.1.2 归一化处理的原因20
3.1.3 归一化算法...20
3.1.4 Matlab 数据归一化处理函数....21
3.2 使用Matlab 实现神经网络.. 21
3.2.1 newff 函数 22
3.2.2 train 函数.. 23
3.2.3 sim 函数.... 23