论文部分内容阅读
Its theoretically desirable to perform variable selection via penalized likelihood that directly penalizes the number of variables in generalized linear models(L0 penalized methods).