摘要
为了克服传统的机器学习方法在供应链管理领域应用存在的局限性,介绍了一种新的支持向量机的机器学习算法。以企业为背景,运用支持向量机算法来解决多类分类问题和函数回归问题。通过在某企业供应链伙伴选择中的实际应用,并与用神经元网络训练得出的结果进行对比,证明这种支持向量机的机器学习算法,不仅具有较高的训练效率,而且有更高的精确度。
To overcome the limitation of the traditional machine learning algorithms in Supply Chain (SC) field, a new machine-learning algorithm of Support Vector Machine(SVM) was proposed. The algorithm differed from those traditional ones in as neural networks. It could resolve the problem of Partner Enterprise Selection in SC Management more efficiently. In order to show its superiority, a real training experiment based on the data from a manufacture Enterprise was discussed in detail. Compared with the results derived from neural networks, the experimental results show that SVM not only improves the training efficiency, but also possesses higher accuracy.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2004年第7期796-800,共5页
Computer Integrated Manufacturing Systems
关键词
供应链
支持向量机
统计学习理论
伙伴企业
supply chain
support vector machine
statistical learning theory
partner enterprise
