PyCM : New statistical analysis library for post classification in Python

A classifier is expected to face with many datasets with different characteristics such as being unbalanced. Besides, their missions are different, for example, categorizing data into just two classes or more than two. There are many different parameters for evaluating the performance of a classifier and each of them evaluates it from a different point of view. Here we want to introduce an open source Python module which evaluates a classifier with many existing basic and professional parameters. This module which named PyCM provides a wide range of overall and classwise evaluating parameters for Python users.
PyCM is a multi-class confusion matrix library written in Python that supports both input data vectors and direct matrix, and a proper tool for post-classification model evaluation that supports most classes and overall statistics parameters. PyCM is the swiss-army knife of confusion matrices, targeted mainly at data scientists that need a broad array of metrics for predictive models and an accurate evaluation of a large variety of classifiers.
In the following links, you can find the source code of this package, many practical examples, and many useful tutorials of it:

Github Repo :
Webpage :


Not a robit
Interesting. I clicked on the link to github. What is the grouping scheme for the metrics, do you collapse the other two groups (columns) as the reference group. It was not intuitive what was going on when results were presented as the three columns per reference group. I would use this if it was in R -> haven't committed to using Python regularly.