文献类型：会议论文

英文题名：A SOM-Based Method for Manifold Learning and Visualization

作者：Shao, Chao[1];Zhang, Xinxiang[1];Wan, Chunhong[1];Shang, Wenqian[2]

第一作者：邵超

通讯作者：Shao, C[1]

机构：[1]Henan Univ Finance & Econ, Sch Informat, Zhengzhou 450002, Peoples R China;[2]Commun Univ China, Sch Comp, Beijing 100024, Peoples R China

第一机构：河南财经政法大学计算机与信息工程学院

通讯机构：[1]corresponding author), Henan Univ Finance & Econ, Sch Informat, Zhengzhou 450002, Peoples R China.|[1048412]河南财经政法大学计算机与信息工程学院;[10484]河南财经政法大学;

会议论文集：2nd International Joint Conference on Computational Sciences and Optimization (CSO)

会议日期：APR 24-26, 2009

会议地点：Sanya, PEOPLES R CHINA

语种：英文

外文关键词：Learning systems - Visualization

摘要：To avoid getting stuck in local minima and obtain better visualization results for data sets lying on low-dimensional nonlinear manifolds embedded in a high-dimensional space, anew SOM-based method, i.e. TO-SOM (Training Orderly-SOM), was presented in this paper. By training the data set orderly according to its neighborhood structure, starting from a small neighborhood in which the data points lie on or close to a locally linear patch, the map can be guided onto the manifold surface and the global visualization results can be achieved step by step. Experimental results show that TO-SOM can discover the intrinsic manifold structure of the data set more faithfully than SOM As a new manifold learning method, TO-SOM is less sensitive to the neighborhood size than other manifold learning methods such as ISOMAP and LLE, which can also be verified by experimental results.