A density and connectivity based decision rule for pattern classification

作者：

Highlights：

• Surrounding Influence Region (SIR) decision rule is proposed for pattern classification.

• SIR decision rule is a parameter-free approach.

• Gabriel Graph is used to define the proximity, connectivity and density relations among the data points.

• A unique neighborhood is constructed for each sample point.

• SIR decision rule is superior to the k-NN and GGN decision rules in artificial and real data sets.

摘要

•Surrounding Influence Region (SIR) decision rule is proposed for pattern classification.•SIR decision rule is a parameter-free approach.•Gabriel Graph is used to define the proximity, connectivity and density relations among the data points.•A unique neighborhood is constructed for each sample point.•SIR decision rule is superior to the k-NN and GGN decision rules in artificial and real data sets.

论文关键词：Classification,Nearest neighbor,Gabriel Graph,Density,Connectivity

论文评审过程：Available online 28 August 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.08.027