Classification Algorithm Based on Feature Selection and Samples Selection
Classification Algorithm Based on Feature Selection and Samples Selection
复制标题
DOI：
10.1007/978-3-642-01510-6_71
复制
发表时间：
2009-05
期刊：
影响因子：
--
通讯作者：
Yitian Xu;Ling Zhen;Liming Yang;Laisheng Wang
中科院分区：
文献类型：
--
作者：
Yitian Xu;Ling Zhen;Liming Yang;Laisheng Wang
提出了一种基于支持向量机和粗糙集理论的分类算法。该方法充分利用粗糙集理论在处理复杂性和不确定性信息方面的优势，首先通过属性约简选择重要特征，然后通过规则归纳选择有效样本，最后利用所选择的重要特征和有效样本构造支持向量分类器。从而降低训练样本的维数，减小训练样本的规模和噪声干扰。它可以为我们提供的好处，提高支持向量机的训练速度和分类精度。图像识别结果验证了该方法的有效性和可行性。为处理大规模高维数据提供了一种有效的方法。
A new classification algorithm based on support vector machine and Rough set theory is proposed in the paper. We make great use of the advantages of Rough set theory in dealing with vagueness and uncertainty information, firstly select important features by attribute reduction; secondly select effective samples by rule induction; finally construct support vector classifier by the selected important features and effective samples. Thus it can reduce training samples’ dimensions, decrease training samples’ scales and noise disturbing. It can provide us with the benefits of improving support vector machine’s training speed and classification accuracy. Result of image recognition verifies its efficiency and feasibility. It also provides us an effective method to deal with the large scale and high dimensions data set.