paul2002 / ml-reporting-tools

Library of functions to generate visually pleasing reports for Machine Learning models

ML-reporting-tools

Library of functions for generating visually pleasing performance reports for Machine Learning models.

Example reports:

Functions:

contains:

• plot_confusion_matrix(): Generates an aesthetic confusion matrix (either standalone or within another figure).
• plot_precision_recall_curve(): Generates an aesthetic Precision-Recall curve (either standalone or within another figure).
• plot_roc_curve(): Generates an aesthetic ROC curve (either standalone or within another figure).
• estimate_best_threshold(): Estimates the "best" threshold for a binary classifier, one that results in the maximum F1 score.
• generate_threshold_report(): Generates a detailed threshold analysis for a binary classifier.
• generate_classification_report(): Generates a detailed report for a classifier.

Contents:

├── source                                  [Directory: Source code]
│   ├── ml_reporting_tools.py               [Main script with all the functions]  
│   ├── example_2_class_classification.py   [Example for a binary classifier]  
│   └── example_3_class_classification.py   [Example for a multiclass classifier]
└── images                                  [Directory: Sample reports]

Dependences

• numpy
• sklearn
• matplotlib