Welcome to Multiverse Tools, a collection of scripts and resources accompanying the manuscript "A sensitivity analysis of preprocessing pipelines: Toward a solution for multiverse analyses".
This repository provides tools and examples to perform sensitivity analyses on preprocessing pipelines. It includes R code for simulations, a real-world application, and a tutorial on how to carry out a sensitivity analysis.
-
tutorial/
R code demonstrating how to compute:- Global effect estimators
- Proportion estimator for pipeline sensitivity analysis
Refer to Section 3 of the manuscript for detailed explanations.
-
sensitivity_analysis-application/
Code (R and Python) for:- Real-world application presented in Section 4.2
Includes data preprocessing and sensitivity analysis steps.
- Real-world application presented in Section 4.2
-
sensitivity_analysis-simulation/
R code used for:- Simulation works discussed in Section 4.1
Covers multiple scenarios to test the proposed sensitivity analysis.
- Simulation works discussed in Section 4.1
Clone this repository:
git clone https://github.com/openneuropet/multiverse_tools.git
Navigate to the desired subdirectory (e.g., tutorial/
) and follow the instructions in the respective script headers.
-
Run the Tutorial
Explore how to run a sensitivity analysis using simulated data.
Navigate totutorial/
and runstatistical_sensitivity_analysis_tutorial.Rmd
. -
Apply to Real Data
Use the code insensitivity_analysis-application/
to replicate the real-world application in Section 4.2. -
Run Simulations
Generate synthetic data and test the sensitivity analysis using the code insensitivity_analysis-simulation/
.
Contributions are welcome! If you identify any issues or have suggestions for improvements, feel free to create a pull request or raise an issue.
This work is licensed under the MIT License.