计算机工程与应用 ›› 2012, Vol. 48 ›› Issue (29): 196-200.
• 图形、图像、模式识别 • 上一篇 下一篇
沈菊红，黄永东，孔妮娜
出版日期:
发布日期:
SHEN Juhong, HUANG Yongdong, KONG Nina
Online:
Published:
摘要： 针对支持向量数据描述中噪声和孤立点带来的过拟合问题，提出了一种Vague集的支持向量数据描述（VFSVDD），利用模糊k-均值聚类方法生成每个训练样本的真、假隶属度，可以精细地控制训练样本对超球面边界的影响。用UCI机器学习数据集的数据实验验证了VFSVDD的有效性。
关键词: 支持向量数据描述, Vague集, 隶属度, 模糊k-均值聚类
Abstract: To resolve the problem of over-fitted caused by noises and outliers in support vector data description, fuzzy support vector data description based on vague sets（VFSVDD） is proposed in this paper. Fuzzy k-means clustering algorithm is employed for generating the truth-membership and false-membership, how each training example affects the boundary of hypersphere could be controlled. Test data from UCI machine learning repository are employed to evaluate the usefulness of VFSVDD.
Key words: Support Vector Data Description（SVDD）, Vague set, membership, fuzzy k-means
沈菊红，黄永东，孔妮娜. 一种Vague集的模糊支持向量数据描述[J]. 计算机工程与应用, 2012, 48(29): 196-200.
SHEN Juhong, HUANG Yongdong, KONG Nina. Fuzzy support vector data description based on vague sets[J]. Computer Engineering and Applications, 2012, 48(29): 196-200.
0 / 推荐
导出引用管理器 EndNote|Ris|BibTeX
链接本文: http://cea.ceaj.org/CN/
http://cea.ceaj.org/CN/Y2012/V48/I29/196
