DATA CLASSIFICATION BASED ON SUPPORTING DATA GRAVITY

摘要：

This paper introduces a novel data classification, method that is based on the idea of data gravity. Many recent clustering and classification ideas based on data gravity tend to consider data gravity magnitude as decisive factor. They eye data gravity as scalar quantity. Novelly in this paper, data gravity is defined to be a vector, and a vector model is set up to classify data by exploiting the internal structure characteristics among vector points in a class. The proposed method is a nonlinear classification technique that can be applied directly on nonlinear separable data sets without concerning nonlinearity-to-linearity transformation (e.g. kernel transformation) of the data. Experiments have showed the validity and some other useful characteristics of this method.

关键词： data classification data gravity nonlinear separable angles between vectors

作者: Li Junlin Fu Hongguang

作者单位: School of Computer Science and Engineering University of Electronic Science and Technology of China ChengDu,China

会议类型: 国际会议

会议名称: 2009 IEEE International Conference on Intelligent Computing and Intelligent Systems(2009 IEEE 智能计算与智能系统国际会议)

会议地点: 上海

会议语种:英文

页码: 22-28

在线出版日期: 2009-11-20（万方平台首次上网日期，不代表论文的发表时间）