Algorithm for Visualization of Classification Results of Two-Category Data
-
摘要: 针对在某些应用领域对二分类数据分类结果可视化的需求,以及现有无监督可视化算法无法提供分类结果的相关信息的问题,提出了二分类数据分类结果可视化算法———支持向量可视化(SVV).该算法是在无监督的自组织神经网络(SOM)的可视化功能的基础上,结合监督学习的支持向量机(SVM)的二分类算法,得到能够直观地显示高维数据、二分类数据分类边界以及数据与分类边界距离的二维映射图,提高了分类结果的可解释性.以SOM可视化算法以及Sammon算法为参照,用2组可分性不同的样本集进行仿真分析,验证了该算法的有效性和可行性.Abstract: A new algorithm called support vector visualization(SVV) was proposed for visualization of classification results of two-category data to meet the need in some applications.The SVV algorithm is based on support vector machine(SVM) and self-organizing mapping(SOM).The result of SVV is a 2D map to visualize highdimensional data,the boundary of the two-category data,as well as the distance between a datum and the boundary.Compared with SOM and Sammon mapping algorithms,experimental results on two datasets with different separability verify the feasibility and effectiveness of the SVV algorithm.
-
Key words:
- SVM /
- SOM /
- visualization /
- algorithm /
- SVV /
- two-category data
-
JACKSON J E.A users guide to principal components[M]. New York:John Wiley and Sons,1991.[2] COX T F,COX M A.Multidimensional scaling[M]. 2nd ed.London:Chapman Hall/CRC Press,2000.[3] KOHONEN T.Self-organizing maps[M]. Berlin:Springer-Verlag,1997:145-152.[4] SAMMON J W.A nonlinear mapping for data structure analysis[J]. IEEE Trans.on Computers,1969(18):401-409.[5] DEMARTINES P,HERAULT J.Curvilinear component analysis:a self-organizing neural network for nonlinear mapping of data sets[J]. IEEE Trans.on Neural Networks,1997(8):148-154.[6] TENENBAUM J B,de SILVA V,LANGFORD J C.A global geometric framework for nonlinear dimensionality reduction[J]. Science,2000(290):2319-2323.[7] LI J X.Visualization of high-dimensional data with relational perspective map[J]. Information Visualization,2004(3):49-59.[8] 邓乃扬,田英杰.数据挖掘中的新方法——支持向量机[M]. 北京:科学出版社,2004:186-206.[9] WU S T,CHOW T W S.Support vector visualization and clustering using self-organizing map and support vector one-class classification[C] ∥Proc.of 2003 Int.Joint Conf.on Neural Networks.Portland:IEEE Press,2003:803-808.[10] SCHLKOPF B,SMOLA A.New support vector algorithms[J]. Neural Computation,2000,12:1 207-1 245.
点击查看大图
计量
- 文章访问数: 1547
- HTML全文浏览量: 91
- PDF下载量: 474
- 被引次数: 0