We consider inference from nonrandom samples in data-rich settings where high-dimensional auxiliary information is available both in the sample and the target population, with survey inference being a special case. We propose a regularized prediction approach ... ...

请注册登录后继续浏览