Interactive classification metrics

Get an intuitive sense for the ROC curve and other binary classification metrics with interactive visualization.

This is a teaching and understanding tool. Change the statistics of the normal distributions or the classification threshold to see how it affects different classification metrics. Read the blog post for more information.

_{^* Matthew's Correlation Coefficient (MCC) represented as unit-normalized MCC as in Cao et al. 2020.}

Install & Run

From PyPI

Create a dedicated python environment (recommended).

python3 -m pip install interactive-classification-metrics

Run with Bokeh server locally from the command line:

run-app

Opens a web browser where you can use the application.

By cloning the repo

1. Clone this repo git clone https://github.com/davhbrown/interactive_classification_metrics.git
2. cd interactive_classification_metrics
3. Create a dedicated python environment is recommended
4. pip install -r requirements.txt

Run with Bokeh server locally from the command line:

bokeh serve --show serve.py

Opens a web browser where you can use the application.

Inspired by

• Cao C, Chicco D, Hoffman MM (2020) The MCC-F1 curve: a performance evaluation technique for binary classification. https://doi.org/10.48550/arXiv.2006.11278
• arthurcgusmao, the author of py-mcc-f1 used here
• Chicco D, Tötsch N, Jurman G (2021) The Matthews correlation coefficient (MCC) is more reliable than balanced accuracy, bookmaker informedness, and markedness in two-class confusion matrix evaluation. BioData Mining 14:13, 1-22.
• Chicco D, Jurman G (2023) The Matthews correlation coefficient (MCC) should replace the ROC AUC as the standard metric for assessing binary classification. BioData Mining 16:4, 1-23.
• the spirit of this tweet

Acknowledgments

Special thanks to Dr. Davide Chicco (@davidechicco) for valuable feedback on this project.