Surrogate-assisted feature extraction for high-throughput phenotyping

This paper introduces a high-throughput unsupervised feature selection method, which improves the robustness and scalability of electronic medical record phenotyping without compromising its accuracy.