基于稀疏聚类的高维数据特征选择及应用

High-Dimension Data Feature Selection and the Application Based on Sparse Clustering

文章研究了一种高维数据聚类特征选择方法——稀疏聚类,稀疏聚类是通过对特征变量赋予权重,并添加lasso惩罚因子,压缩权重,得到对变量的权重排序,即重要性排序,使其在进行分类预测的同时达到自动剔除冗余变量的效果,从而起到了对高维数据聚类时的特征选择作用。将此方法运用于中国环保问题,将中国31个省份根据环保情况分为3类,并从现有的104个环保指标中筛选得到20个重要指标。

This paper investigates a feature selection method in high-dimensional data clustering called sparse clustering,which gives different weights to the features and select the features using a lasso-type penalty. Sparse clustering can help one cluster the observations using an adaptively chosen subset of the features, namely, it plays a role of feature selection in high-dimensional data. this paper applies sparse clustering on China＇ s environmental problems, dividing 31 provinces to 3 parts according to environmental condition and selecting 20 important items from 104 environmental items.