"polysr: Symbolic Regression applied to Polymeric Use Cases" is a GitHub repository that demonstrates the utility of symbolic regression for understanding polymer data. Specifically, three distinct use cases are considered. The Robust to Noise use case demonstrates the robustness of symbolic regression to noise. Both the case where noise is added to the predicted quantity and the case where noise is added to one of the features are considered. Specifically, synthetic data is generated from the Flory Huggins spinodal equation for the tests. The Leveraging Complexity use case demonstrates the power of symbolic regression to learn an equation where the constants depend on the class. Specifically, the Williams–Landel–Ferry equation is learned directly from experimental data where the constants are chemistry dependent. Finally, the Increasing Interpretability use case demonstrates the power of using symbolic regression to find a simple model that is interpretable and combining it with a traditional machine learning model to get accuracy. This approach also provides insight into when the simple model may fail. Specifically, the adsoprtion free energy of sequence defined polymers is used.
About this Dataset
| Title | polysr: use cases demonstrating the utility of symbolic regression for polymeric data |
|---|---|
| Description | "polysr: Symbolic Regression applied to Polymeric Use Cases" is a GitHub repository that demonstrates the utility of symbolic regression for understanding polymer data. Specifically, three distinct use cases are considered. The Robust to Noise use case demonstrates the robustness of symbolic regression to noise. Both the case where noise is added to the predicted quantity and the case where noise is added to one of the features are considered. Specifically, synthetic data is generated from the Flory Huggins spinodal equation for the tests. The Leveraging Complexity use case demonstrates the power of symbolic regression to learn an equation where the constants depend on the class. Specifically, the Williams–Landel–Ferry equation is learned directly from experimental data where the constants are chemistry dependent. Finally, the Increasing Interpretability use case demonstrates the power of using symbolic regression to find a simple model that is interpretable and combining it with a traditional machine learning model to get accuracy. This approach also provides insight into when the simple model may fail. Specifically, the adsoprtion free energy of sequence defined polymers is used. |
| Modified | 2026-03-26 00:00:00 |
| Publisher Name | National Institute of Standards and Technology |
| Contact | mailto:[email protected] |
| Keywords | polymers , machine learning , symbolic regression |
{
"identifier": "ark:\/88434\/mds2-4159",
"accessLevel": "public",
"contactPoint": {
"hasEmail": "mailto:[email protected]",
"fn": "Debra Audus"
},
"programCode": [
"006:045"
],
"landingPage": "",
"title": "polysr: use cases demonstrating the utility of symbolic regression for polymeric data",
"description": "\"polysr: Symbolic Regression applied to Polymeric Use Cases\" is a GitHub repository that demonstrates the utility of symbolic regression for understanding polymer data. Specifically, three distinct use cases are considered. The Robust to Noise use case demonstrates the robustness of symbolic regression to noise. Both the case where noise is added to the predicted quantity and the case where noise is added to one of the features are considered. Specifically, synthetic data is generated from the Flory Huggins spinodal equation for the tests. The Leveraging Complexity use case demonstrates the power of symbolic regression to learn an equation where the constants depend on the class. Specifically, the Williams\u2013Landel\u2013Ferry equation is learned directly from experimental data where the constants are chemistry dependent. Finally, the Increasing Interpretability use case demonstrates the power of using symbolic regression to find a simple model that is interpretable and combining it with a traditional machine learning model to get accuracy. This approach also provides insight into when the simple model may fail. Specifically, the adsoprtion free energy of sequence defined polymers is used.",
"language": [
"en"
],
"distribution": [
{
"accessURL": "https:\/\/github.com\/usnistgov\/polysr",
"description": "A GitHub repository that contains use cases demonstrating the utility of symbolic regression as applied to polymeric data.",
"title": "polysr"
}
],
"bureauCode": [
"006:55"
],
"modified": "2026-03-26 00:00:00",
"publisher": {
"@type": "org:Organization",
"name": "National Institute of Standards and Technology"
},
"theme": [
"Materials:Polymers",
"Information Technology:Data and informatics"
],
"keyword": [
"polymers",
"machine learning",
"symbolic regression"
]
}
