ppernot / ErrView

Error sets viewer and comparator

ErrView: Error sets viewer and comparator

Implements methods from the following articles:

1. P. Pernot and A. Savin (2018) Probabilistic performance estimators for computational chemistry methods: The empirical cumulative distribution function of absolute errors. J. Chem. Phys. 148:241707. https://doi.org/10.1063/1.5016248

2. P. Pernot and A. Savin (2020) Probabilistic performance estimators for computational chemistry methods: Systematic Improvement Probability and Ranking Probability Matrix. I. Theory. J. Chem. Phys. 152:164108. http://dx.doi.org/10.1063/5.0006202

3. P. Pernot and A. Savin (2021) Using the Gini coefficient to characterize the shape of computational chemistry error distributions. Theor. Chem. Acc. 140:24 https://doi.org/10.1007/s00214-021-02725-0 ; https://arxiv.org/abs/2012.09589

Please cite these articles if you use the code.

Data

Several datasets treated in the article are available here. To use your own data, create a .csv file with the following columns:

• "" : systems names or codes
• "Ref": reference data
• "M1" : results of first method (replace M1 by the actual name)
• ...
• "Mn" : results of the last method

Run it on the cloud

1. Click on the binder badge above: this will open a Rstudio session in your browser

2. Select server.R in the Files panel of Rstudio

3. Click on Run App

4. Enjoy !

If you want to use the predefined datasets, you will have to download them first from here.

Test before install

A web interface to the code is also available online here.

Predefined datasets are availble directly on the server.

Warning ! This version runs on a server with limited memory. It will handle only small datasets. Running out of memory will generate a message "Error: An error has occurred. Check your logs or contact the app author for clarification." Sorry, there is nothing the author can do for you. Please try the binder method (see above) of install the code.