Abstract:
This paper introduces a new machine learning method——support vector machine (SVM) and analyzes some of the characteristics of SVM. After linearly scaling and principal components analyzing (PCA) the data that affect product’s quality, it trains them by the tools of Libsvm. Compared to Bayesian theory, BP neural network and decision tree(ID3), the result of SVM is excellent. Based on the models, the production of processing industry can be improved.
Key words:
SVM,
Statistical learning theory,
Monitor product quality,
Process industry
摘要: 采用了一种新的机器学习方法——支持向量机,对流程企业中的历史数据进行分类分析,阐述了支持向量机的基本内容,对影响产品质量因素的样本集进行了标准化处理和主因素分析(PCA),采用Libsvm训练了数据集,并与贝叶斯理论、BP神经网络和决策树ID3的分类结果比较,证明了算法的优越性,为产品的质量监控提供了有效依据。
关键词:
支持向量机,
统计学习理论,
质量监控,
流程企业
YAN Wei; ZHANG Hao; LU Jianfeng. Study and Application of Support Vector Machine to Monitor Product Quality[J]. Computer Engineering, 2006, 32(15): 184-186.
闫 伟;张 浩;陆剑峰. 基于支撑向量机的产品质量监控研究应用[J]. 计算机工程, 2006, 32(15): 184-186.