摘要
对于已经分类的数据和大量未分类数据,在运算过程中,采用一种新的半监督聚类算法为支持向量机提供新的训练数据.随后,利用支持向量机判别出所有数据的类别属性,并选取最可靠的点加入已分类集合.为了验证算法的效率,收集了67张黄瓜叶片色调的数字信息,并对具有6个已分类数据与61个未分类数据的数据集进行半监督聚类分析,以判断这些叶片的健康程度.结果表明,该聚类算法优于其他算法.
Durning the run of this algorithm, a novel Fuzzy C-Means Algorithm is used to provide a better training set for a SVM;while a SVM is used to determine the class labels of the whole data set, and the most reliable data is then added to the labeled set. To test the performance of our algorithm, 67 pieces of cucumber leaves which are infected with downy mildew are classified, The result shows this kind of algorithm is better than some other semi-supervised kinds.
出处
《泉州师范学院学报》
2013年第6期49-52,共4页
Journal of Quanzhou Normal University
基金
泉州市科技计划项目(2012Z103)
福建省教育厅科技项目(JK2013037
JA12273)
