Estimation of Intrinsic Dimension based on Denseness and Dispersal

摘要：

We propose a new algorithm to estimate the intrinsic dimension of data sets. The method is reconstructed by the tradeoff analysis on denseness and dispersal, which come from the intrinsic characters of data set. At last, this method is compared to principal component analysis and the correlation dimension. Experiments show that this method is more reasonable and more reliable.

关键词： Denseness Dispersal Intrinsic Dimension Tradeoff Analysis

作者: TAN Lu CHEN Xiaolan GUO Hongfeng

作者单位: Dept of Statistics and Mathematics of Shandong University of finance, Jinan, P.R.China, 250014

会议类型: 国际会议

会议名称: 2009 International Institute of Applied Statistics Studies(2009 国际应用统计学术研讨会)

会议地点: 青岛

会议语种:英文

页码: 1-6

在线出版日期: 2009-07-25（万方平台首次上网日期，不代表论文的发表时间）