SHAP explainer for Case Study A.

Gespeichert in:
Bibliographische Detailangaben
Titel: SHAP explainer for Case Study A.
Autoren: Nathan C. Hurley, Nihar Desai, Sanket S. Dhruva, Rohan Khera, Wade Schulz, Chenxi Huang, Jeptha Curtis, Frederick Masoudi, John Rumsfeld, Sahand Negahban, Harlan M. Krumholz, Bobak J. Mortazavi
Publikationsjahr: 2025
Schlagwörter: Medicine, Biotechnology, Science Policy, Mental Health, Hematology, Infectious Diseases, past medical history, key decision points, individualized care throughout, experienced bleeds ), 712 patients classified, percutaneous coronary intervention, closure method categories, update risk prediction, date risk prediction, date risk information, 170 pcis performed, newly available data, using data available, 3 %) remaining, estimate evolving risk, data available, laboratory data, coronary anatomy, closure device, 7 %), risk requires, moderate risk, low risk, index pcis
Beschreibung: At each model stage, the prediction is created by summing each variable contribution to risk. Variables on the left (red) are contribute to an increased risk of bleeding, while variables on the right (blue) contribute to a decreased risk of bleeding. Variables are organized such that those providing the strongest change to risk are at the center, with variables providing smaller changes to risk at the outside. (DOCX)
Publikationsart: article in journal/newspaper
Sprache: unknown
Relation: https://figshare.com/articles/journal_contribution/SHAP_explainer_for_Case_Study_A_/29404541
DOI: 10.1371/journal.pdig.0000906.s025
Verfügbarkeit: https://doi.org/10.1371/journal.pdig.0000906.s025
https://figshare.com/articles/journal_contribution/SHAP_explainer_for_Case_Study_A_/29404541
Rights: CC BY 4.0
Dokumentencode: edsbas.5717F4A4
Datenbank: BASE
Beschreibung
Abstract:At each model stage, the prediction is created by summing each variable contribution to risk. Variables on the left (red) are contribute to an increased risk of bleeding, while variables on the right (blue) contribute to a decreased risk of bleeding. Variables are organized such that those providing the strongest change to risk are at the center, with variables providing smaller changes to risk at the outside. (DOCX)
DOI:10.1371/journal.pdig.0000906.s025