Applying active learning to high-throughput phenotyping algorithms for electronic health records data.

By Yukun Chen, Robert J. Carroll, E. M. Hinz, Anushi Shah, A. Eyler, Joshua Charles Denny, Hua Xu. 2013

115 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Health Information Management

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Jimeng Sun
2. Yuan Luo
3. Zexian Zeng
4. Tristan Naumann
5. Walter F. Stewart
6. Xiaoyu Li
7. Yu Deng
8. Ozlem Uzuner

In Health Information Management

1. Yuan Luo
2. Jimeng Sun
3. Ozlem Uzuner
4. Xiaoyu Li
5. Yu Deng
6. Zexian Zeng
7. Joshua Charles Denny
8. Bradley A. Malin

Citations in the corpus

• Natural Language Processing for EHR-Based Computational Phenotyping (2018, methodology): “[99] applied active learning to a SVM classification algorithm to identify rheumatoid, colorectal cancer, and venous thromboembolism.”
• Identifying and Mitigating Biases in EHR Laboratory Tests (2014, unknown): “The work presented in this paper is highly relevant to researchers working on cohort identification algorithms, especially with the recent push for more automated ways to perform high-throughput phenotyping [34–36].”
• Limestone: High-throughput candidate phenotype generation via tensor factorization (2014, methodology): “Yet, the current state of the art high-throughput phenotyping cannot generate large amounts of candidate phenotypes that simultaneously achieve good performance without human annotated samples [37].”
• A systematic comparison of feature space effects on disease classifier performance for phenotype identification of five diseases (2015, methodology): “…Eyler, & Denny, 2011; Chen et al., 2013), diabetes (Wright, McCoy, Henkin, Kale, & Sittig, 2013), colorectal cancer, and venous thromboembolism (Chen et al., 2013), in risk adjustment for ICU patients (Marafino, Davies, Bardach, Dean, & Dudley, 2014) and for smoking history detection (Khor et…”
• Large-Scale Discovery of Disease-Disease and Disease-Gene Associations (2016, background): “Some researchers proposed to combat this issue by employing active learning approaches to obtain limited number of expert labels used by supervised methods 23,24 .”
• Rubik: Knowledge Guided Tensor Factorization and Completion for Health Data Analytics (2015, methodology): “Most existing phenotyping methods are supervised approaches, which are either expert-defined rule-based meth-ods [16] or classification methods [8].”
• Marble: high-throughput phenotyping from electronic health records via sparse nonnegative tensor factorization (2014, methodology): “Although data mining tools have been utilized to automate the phenotype process, current high-throughput methodologies cannot generate large amounts of candidate phenotypes and achieve good performance without human annotated samples [5].”