Literature DB >> 35426912

Benchmarking missing-values approaches for predictive models on health databases.

Alexandre Perez-Lebel1,2,3, Gaël Varoquaux1,2,3, Marine Le Morvan2, Julie Josse4,5, Jean-Baptiste Poline1.   

Abstract

BACKGROUND: As databases grow larger, it becomes harder to fully control their collection, and they frequently come with missing values. These large databases are well suited to train machine learning models, e.g., for forecasting or to extract biomarkers in biomedical settings. Such predictive approaches can use discriminative-rather than generative-modeling and thus open the door to new missing-values strategies. Yet existing empirical evaluations of strategies to handle missing values have focused on inferential statistics.
RESULTS: Here we conduct a systematic benchmark of missing-values strategies in predictive models with a focus on large health databases: 4 electronic health record datasets, 1 population brain imaging database, 1 health survey, and 2 intensive care surveys. Using gradient-boosted trees, we compare native support for missing values with simple and state-of-the-art imputation prior to learning. We investigate prediction accuracy and computational time. For prediction after imputation, we find that adding an indicator to express which values have been imputed is important, suggesting that the data are missing not at random. Elaborate missing-values imputation can improve prediction compared to simple strategies but requires longer computational time on large data. Learning trees that model missing values-with missing incorporated attribute-leads to robust, fast, and well-performing predictive modeling.
CONCLUSIONS: Native support for missing values in supervised machine learning predicts better than state-of-the-art imputation with much less computational cost. When using imputation, it is important to add indicator columns expressing which values have been imputed.
© The Author(s) 2022. Published by Oxford University Press GigaScience.

Entities:  

Keywords:  bagging; benchmark; imputation; machine learning; missing values; multiple imputation; supervised learning

Mesh:

Year:  2022        PMID: 35426912      PMCID: PMC9012100          DOI: 10.1093/gigascience/giac013

Source DB:  PubMed          Journal:  Gigascience        ISSN: 2047-217X            Impact factor:   7.658


  17 in total

1.  Missing data should be handled differently for prediction than for description or causal explanation.

Authors:  Matthew Sperrin; Glen P Martin; Rose Sisk; Niels Peek
Journal:  J Clin Epidemiol       Date:  2020-06-12       Impact factor: 6.437

2.  Machine learning for causal inference in Biostatistics.

Authors:  Sherri Rose; Dimitris Rizopoulos
Journal:  Biostatistics       Date:  2020-04-01       Impact factor: 5.899

3.  Benchmarking missing-values approaches for predictive models on health databases.

Authors:  Alexandre Perez-Lebel; Gaël Varoquaux; Marine Le Morvan; Julie Josse; Jean-Baptiste Poline
Journal:  Gigascience       Date:  2022-04-15       Impact factor: 7.658

4.  Data-driven advice for applying machine learning to bioinformatics problems.

Authors:  Randal S Olson; William La Cava; Zairah Mustahsan; Akshay Varik; Jason H Moore
Journal:  Pac Symp Biocomput       Date:  2018

5.  Strategies for handling missing data in electronic health record derived data.

Authors:  Brian J Wells; Kevin M Chagin; Amy S Nowacki; Michael W Kattan
Journal:  EGEMS (Wash DC)       Date:  2013-12-17

6.  Deep Patient: An Unsupervised Representation to Predict the Future of Patients from the Electronic Health Records.

Authors:  Riccardo Miotto; Li Li; Brian A Kidd; Joel T Dudley
Journal:  Sci Rep       Date:  2016-05-17       Impact factor: 4.379

7.  Polygenic prediction of breast cancer: comparison of genetic predictors and implications for risk stratification.

Authors:  Kristi Läll; Maarja Lepamets; Marili Palover; Tõnu Esko; Andres Metspalu; Neeme Tõnisson; Peeter Padrik; Reedik Mägi; Krista Fischer
Journal:  BMC Cancer       Date:  2019-06-10       Impact factor: 4.430

8.  MIMIC-III, a freely accessible critical care database.

Authors:  Alistair E W Johnson; Tom J Pollard; Lu Shen; Li-Wei H Lehman; Mengling Feng; Mohammad Ghassemi; Benjamin Moody; Peter Szolovits; Leo Anthony Celi; Roger G Mark
Journal:  Sci Data       Date:  2016-05-24       Impact factor: 6.444

9.  Biases in electronic health record data due to processes within the healthcare system: retrospective observational study.

Authors:  Denis Agniel; Isaac S Kohane; Griffin M Weber
Journal:  BMJ       Date:  2018-04-30

10.  Machine learning models in electronic health records can outperform conventional survival models for predicting patient mortality in coronary artery disease.

Authors:  Andrew J Steele; Spiros C Denaxas; Anoop D Shah; Harry Hemingway; Nicholas M Luscombe
Journal:  PLoS One       Date:  2018-08-31       Impact factor: 3.240

View more
  1 in total

1.  Benchmarking missing-values approaches for predictive models on health databases.

Authors:  Alexandre Perez-Lebel; Gaël Varoquaux; Marine Le Morvan; Julie Josse; Jean-Baptiste Poline
Journal:  Gigascience       Date:  2022-04-15       Impact factor: 7.658

  1 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.