Double/debiased machine learning for treatment and structural parameters
University of California, Los Angeles · Massachusetts Institute of Technology · +2 more institutions
Abstract
We revisit the classic semiparametric problem of inference on a low di-mensional parameter Ø0 in the presence of high-dimensional nuisance parameters Û0. We depart from the classical setting by allowing for Û0 to be so high-dimensional that the traditional assumptions, such as Donsker properties, that limit complexity of the parameter space for this object break down. To estimate Û0, we consider the use of statistical or machine learning (ML) methods which are particularly well-suited to estimation in modern, very high-dimensional cases. ML methods perform well by employing regularization to reduce variance and trading off regularization bias with overfitting in practice. However, both regularization bias and…
Citation impact
- FWCI
- —
- Percentile
- —
- References
- 105
Authors
7- VCVictor ChernozhukovCorresponding
University of California, Los Angeles, Massachusetts Institute of Technology
- DCDenis Chetverikov
University of Chicago
- MDMert Demirer
University of California, Los Angeles, Massachusetts Institute of Technology
- EDEsther Duflo
Harvard University
- CHChristian Hansen
University of California, Los Angeles, Massachusetts Institute of Technology
Topics & keywords
- Computer science
- Artificial intelligence
- Machine learning
- Pattern recognition (psychology)