docs/index.md

In ppernot/ErrView: Analysis of error sets

ErrView User's Manual

(2021-21-06)

Introduction

Errview provides a GUI to analyze error datasets and perform computational chemistry methods benchmarking and ranking.

It is mostly based on the papers by P. Pernot and A. Savin [1-4], where a complete description of the methods and further links to the literature are provided.

The entry page provides a short help and a set of links to various pages, including this one.

The only mandatory module is the first one (Data), the other ones implement different analysis methods.

• Data: load a dataset

• Outliers: identify global outliers

• Statistics: calculate statistics

• HistDist: plot the errors and their histograms

• QQ-Plot: qq-plots

• PDFs: plot probability density functions

• ECDFs: plot the empirical cumulated distribution functions

• Lorenz: lorenz curves

• Correlations: correlation matrices of data, errors and statistics

• Ranking: probabilistic ranking matrix

• SIP Mat: systematic improvement probability matrix

• Delta |Err.|: SIP analysis of fifferences of absolute errors

• UQ: uncertainty quantification (Work in Progress)

References

1. Pernot and A. Savin (2018) J. Chem. Phys. 148:241707

2. P. Pernot and A. Savin (2020) J. Chem. Phys. 152:164108; arXiv

3. P. Pernot and A. Savin (2020) J. Chem. Phys. 152:164109; arXiv

4. P. Pernot and A. Savin (2021) arXiv

ppernot/ErrView documentation built on Jan. 30, 2022, 6:59 a.m.