一种改进的HyperMap可视化降维方法  

An Improved HyperMap Dimension Reduction Method for Visualization

作　　者：郭崇慧[1] 李丽薇[1] 

机构地区：[1]大连理工大学系统工程研究所,大连116024

出　　处：《情报学报》2012年第10期1077-1082,共6页Journal of the China Society for Scientific and Technical Information

基　　金：国家自然科学基金资助项目（71031002,71171030）;新世纪优秀人才支持计划资助（NCET-11-0050）

摘　　要：可视化方法可应用于知识发现的各个阶段。降维是知识发现可视化过程中处理多维数据的一种常用方法。本文基于FastMap与HyperMap方法，提出一种改进的HyperMap可视化降维方法。该方法首先选择数据集中相距最远的样本点作为枢纽点，然后根据应力函数建立优化模型并进行优化求解，以确定最优的权重组合，并允许用户对可视化结果进行旋转。数值实验结果验证了该方法的可行性与有效性。Visualization is often used in different knowledge discovery phrases and dimension reduction is a frequently-used method to deal with multi-dimensional data in visualization process. Based on FastMap and HyperMap dimension reduction methods, this paper provides an improved HyperMap dimension reduction method. Firstly, this method chooses the farthest points in data set as pivots and then establishes optimization model according to stress function. By way of solving optimization model, it searches the best weights. In addition, this method allows users to rotate visualization results. Experimental results validate its feasibility and effectiveness.

关 键 词：可视化方法 降维 FastMap HyperMap枢纽点 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]