Data for: A systematic review showed no performance benefit of machine learning over logistic regression for clinical prediction models

Published: 14-03-2019| Version 1 | DOI: 10.17632/sypyt6c2mc.1
Ben Van Calster,
Jie Ma,
Evangelia Christodoulou,
Ewout Steyerberg,
Jan Verbakel,
Gary Collins


The uploaded files are: 1) Excel file containing 6 sheets in respective Order: "Data Extraction" (summarized final data extractions from the three reviewers involved), "Comparison Data" (data related to the comparisons investigated), "Paper level data" (summaries at paper level), "Outcome Event Data" (information with respect to number of events for every outcome investigated within a paper), "Tuning Classification" (data related to the manner of hyperparameter tuning of Machine Learning Algorithms). 2) R script used for the Analysis (In order to read the data, please: Save "Comparison Data", "Paper level data", "Outcome Event Data" Excel sheets as txt files. In the R script srpap: Refers to the "Paper level data" sheet, srevents: Refers to the "Outcome Event Data" sheet and srcompx: Refers to " Comparison data Sheet". 3) Supplementary Material: Including Search String, Tables of data, Figures 4) PRISMA checklist items