Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure
Abstract Machine learning is increasingly being used to predict clinical outcomes. Most comparisons of different methods have been based on empirical analyses in specific datasets. We used Monte Carlo simulations to determine when machine learning methods perform better than statistical learning met...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2022-06-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-022-13015-5 |
| _version_ | 1828724937514287104 |
|---|---|
| author | Peter C. Austin Frank E. Harrell Douglas S. Lee Ewout W. Steyerberg |
| author_facet | Peter C. Austin Frank E. Harrell Douglas S. Lee Ewout W. Steyerberg |
| author_sort | Peter C. Austin |
| collection | DOAJ |
| description | Abstract Machine learning is increasingly being used to predict clinical outcomes. Most comparisons of different methods have been based on empirical analyses in specific datasets. We used Monte Carlo simulations to determine when machine learning methods perform better than statistical learning methods in a specific setting. We evaluated six learning methods: stochastic gradient boosting machines using trees as the base learners, random forests, artificial neural networks, the lasso, ridge regression, and linear regression estimated using ordinary least squares (OLS). Our simulations were informed by empirical analyses in patients with acute myocardial infarction (AMI) and congestive heart failure (CHF) and used six data-generating processes, each based on one of the six learning methods, to simulate continuous outcomes in the derivation and validation samples. The outcome was systolic blood pressure at hospital discharge, a continuous outcome. We applied the six learning methods in each of the simulated derivation samples and evaluated performance in the simulated validation samples. The primary observation was that neural networks tended to result in estimates with worse predictive accuracy than the other five methods in both disease samples and across all six data-generating processes. Boosted trees and OLS regression tended to perform well across a range of scenarios. |
| first_indexed | 2024-04-12T13:18:28Z |
| format | Article |
| id | doaj.art-5ea96a4f98254d9581c5a4db9ccbd416 |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-12T13:18:28Z |
| publishDate | 2022-06-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-5ea96a4f98254d9581c5a4db9ccbd4162022-12-22T03:31:36ZengNature PortfolioScientific Reports2045-23222022-06-0112111110.1038/s41598-022-13015-5Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressurePeter C. Austin0Frank E. Harrell1Douglas S. Lee2Ewout W. Steyerberg3ICESDepartment of Biostatistics, Vanderbilt University School of MedicineICESDepartment of Biomedical Data Sciences, Leiden University Medical CentreAbstract Machine learning is increasingly being used to predict clinical outcomes. Most comparisons of different methods have been based on empirical analyses in specific datasets. We used Monte Carlo simulations to determine when machine learning methods perform better than statistical learning methods in a specific setting. We evaluated six learning methods: stochastic gradient boosting machines using trees as the base learners, random forests, artificial neural networks, the lasso, ridge regression, and linear regression estimated using ordinary least squares (OLS). Our simulations were informed by empirical analyses in patients with acute myocardial infarction (AMI) and congestive heart failure (CHF) and used six data-generating processes, each based on one of the six learning methods, to simulate continuous outcomes in the derivation and validation samples. The outcome was systolic blood pressure at hospital discharge, a continuous outcome. We applied the six learning methods in each of the simulated derivation samples and evaluated performance in the simulated validation samples. The primary observation was that neural networks tended to result in estimates with worse predictive accuracy than the other five methods in both disease samples and across all six data-generating processes. Boosted trees and OLS regression tended to perform well across a range of scenarios.https://doi.org/10.1038/s41598-022-13015-5 |
| spellingShingle | Peter C. Austin Frank E. Harrell Douglas S. Lee Ewout W. Steyerberg Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure Scientific Reports |
| title | Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| title_full | Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| title_fullStr | Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| title_full_unstemmed | Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| title_short | Empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| title_sort | empirical analyses and simulations showed that different machine and statistical learning methods had differing performance for predicting blood pressure |
| url | https://doi.org/10.1038/s41598-022-13015-5 |
| work_keys_str_mv | AT petercaustin empiricalanalysesandsimulationsshowedthatdifferentmachineandstatisticallearningmethodshaddifferingperformanceforpredictingbloodpressure AT frankeharrell empiricalanalysesandsimulationsshowedthatdifferentmachineandstatisticallearningmethodshaddifferingperformanceforpredictingbloodpressure AT douglasslee empiricalanalysesandsimulationsshowedthatdifferentmachineandstatisticallearningmethodshaddifferingperformanceforpredictingbloodpressure AT ewoutwsteyerberg empiricalanalysesandsimulationsshowedthatdifferentmachineandstatisticallearningmethodshaddifferingperformanceforpredictingbloodpressure |