PWSHAP: A Path-Wise Explanation Model for Targeted Variables

Authors: Lucile Ter-Minassian, Oscar Clivio, Karla Diazordaz, Robin J. Evans, Christopher C. Holmes

ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the resolution, interpretability and true locality of our approach on examples and a real-world experiment.
Researcher Affiliation Academia 1Department of Statistics, University of Oxford, Oxford, UK 2Department of Statistical Science, University College London, London, UK 3The Alan Turing Institute, London, UK.
Pseudocode No The paper does not contain any clearly labeled pseudocode or algorithm blocks.
Open Source Code Yes The code used for experiments is available at https://github.com/oscarclivio/pwshap.
Open Datasets Yes We present a local mediation analysis experiment on the Adult data set from UCI (Asuncion & Newman, 2007)... The URL for this dataset is https://archive.ics.uci.edu/ml/datasets/adult.
Dataset Splits Yes The 5000 observations in the set were subsampled again 10 times, by taking out 20% of the sample. Each time, we use the subsample as both a training and testing set for our models.
Hardware Specification Yes Experiments were ran using a 2,6 GHz 6-Core Intel Core i7.
Software Dependencies No The paper mentions software like “sklearn package” but does not provide specific version numbers for these dependencies.
Experiment Setup No The paper describes aspects of the experimental setup, such as modeling choices (e.g., “linear and logistic regression”), but does not provide specific hyperparameter values or detailed training configurations in the main text.