高维数据的特征选择研究

Some Studies on Feature Selection for High Dimensional Data

特征选择是机器学习的重要研究内容之一.相对于低维数据的特征选择而言,高维数据的特征选择更具挑战性,尤其是高维小样本的特征选择问题,因而吸引很多研究者的关注.高维特征选择问题称为稀疏建模问题,其目标是解决现有特征建模方法在高维特征空间失效的问题.本文对高维数据的特征选择研究成果进行了相应的总结和展望.

Feature selection is a key issue in machine learning field.As compared with feature selection for low dimensional data,feature selection for high dimensional data is a challenging task,especially feature selection issue for high dimensional small size data,so many researcher focus on this problem.In essence,the feature selection problem for high dimensional data is regarded as a sparse modeling issue,whose target is to solve the failure problem of the existing feature modeling methods on high dimensional feature space.Therefore,in this paper,we give a survey of the feature selection methods for high dimensional data,and meanwhile propose some discussions on future work.Our main objective is to provide a reference for readers who are interesting in this research field.