：Lasso 是一种基于一范式的特征选择方法。与已有的特征选择方法相比较，Lasso 不仅能够准确地选择出与类标签强相

关的变量，同时还具有特征选择的稳定性，因而成为人们研究的一个热点-: Lasso is a paradigm based on a feature selection method. Compared with the existing feature selection methods, Lasso not only to accurately select a strong correlation with the class label variables, but also has the stability of feature selection, and thus become a hot research