#sklearn-evaluation

scikit-learn model evaluation made easy: plots, tables and markdown reports.

Works with Python 2 and 3.

Documentation here.

Install

pip install sklearn-evaluation

Optional dependencies

If you want to use the reports module you need to install mistune and tabulate for the tables module

# for generating reports
pip install mistune
# for the table module
pip install tabulate

#Usage

##plot module

Generate evaluation plots with a single function call.

from sklearn_evaluation import plot

#code for data loading and model training

plot.confusion_matrix(y_true, y_pred, target_names=target_names)

##table module

Generate good looking tables from your model results.

from sklearn_evaluation import table

#code for data loading and training

table.feature_importances(model)

+-----------+--------------+-----------+
| name      |   importance |       std |
+===========+==============+===========+
| Feature 0 |    0.250398  | 0.0530907 |
+-----------+--------------+-----------+
| Feature 1 |    0.232397  | 0.0523836 |
+-----------+--------------+-----------+
| Feature 2 |    0.148898  | 0.0331814 |
+-----------+--------------+-----------+
| Feature 3 |    0.0553634 | 0.0128296 |
+-----------+--------------+-----------+
| Feature 8 |    0.05401   | 0.0122248 |
+-----------+--------------+-----------+
| Feature 5 |    0.053878  | 0.01289   |
+-----------+--------------+-----------+
| Feature 6 |    0.0525828 | 0.0130225 |
+-----------+--------------+-----------+
| Feature 9 |    0.0510197 | 0.0129436 |
+-----------+--------------+-----------+
| Feature 7 |    0.0509633 | 0.0117197 |
+-----------+--------------+-----------+
| Feature 4 |    0.0504887 | 0.012844  |
+-----------+--------------+-----------+

Also, running this in Jupyter will generate a pandas-like output.

##Using the OOP interface

A simplified API is available by packing the results of your estimator in the ClassifierEvaluator class.

from sklearn_evaluation import ClassifierEvaluator

# code for data loading and model training

ce = ClassifierEvaluator(classifier, y_test, y_pred, y_score,
                         feature_list, target_names)

# this plots the confusion matrix
ce.confusion_matrix

Generating reports

Generate reports using Markdown templates.

 
 template = '''
            # Report
            {estimator_type}
            {date}
            {confusion_matrix}
            {roc}
            {precision_recall}
            '''

ce.generate_report(template, path='report.html')

The code above will generate a report like this one.

Reports are self-contained, all images are included in the html file using base64.