Metadata-Version: 2.1
Name: clana
Version: 0.3.2
Summary: CLANA is a toolkit for classifier analysis.
Home-page: https://github.com/MartinThoma/clana
Author: Martin Thoma
Author-email: info@martin-thoma.de
Maintainer: Martin Thoma
Maintainer-email: info@martin-thoma.de
License: MIT
Download-URL: https://github.com/MartinThoma/language-identification
Description: [![DOI](https://zenodo.org/badge/102892750.svg)](https://zenodo.org/badge/latestdoi/102892750)
        [![PyPI version](https://badge.fury.io/py/clana.svg)](https://badge.fury.io/py/clana)
        [![Python Support](https://img.shields.io/pypi/pyversions/clana.svg)](https://pypi.org/project/clana/)
        [![Documentation Status](https://readthedocs.org/projects/clana/badge/?version=latest)](http://clana.readthedocs.io/en/latest/?badge=latest)
        [![Build Status](https://travis-ci.org/MartinThoma/clana.svg?branch=master)](https://travis-ci.org/MartinThoma/clana)
        [![Coverage Status](https://coveralls.io/repos/github/MartinThoma/clana/badge.svg?branch=master)](https://coveralls.io/github/MartinThoma/clana?branch=master)
        
        # clana
        
        `clana` is a toolkit for classifier analysis. One key contribution of clana is
        Confusion Matrix Ordering (CMO) as explained in chapter 5 of [Analysis and Optimization of Convolutional Neural Network Architectures](https://arxiv.org/abs/1707.09725). It is a technique
        that can be applied to any multi-class classifier and helps to understand which
        groups of classes are most similar.
        
        
        ## Installation
        
        The recommended way to install clana is:
        
        ```
        $ pip install clana --user
        ```
        
        If you want the latest version:
        
        ```
        $ git clone https://github.com/MartinThoma/clana.git; cd clana
        $ pip install -e . --user
        ```
        
        ## Usage
        
        ```
        clana --help
        Usage: clana [OPTIONS] COMMAND [ARGS]...
        
        Options:
          --version  Show the version and exit.
          --help     Show this message and exit.
        
        Commands:
          distribution   Get the distribution of classes in a dataset.
          get-cm         Calculate the confusion matrix (CSV inputs).
          get-cm-simple  Calculate the confusion matrix (one label per...
          visualize      Optimize confusion matrix.
        ```
        
        The visualize command gives you images like this:
        
        ![Confusion Matrix after Confusion Matrix Ordering of the WiLI-2018 dataset](cm-wili-2018.png)
        
        
        ### Data distribution
        
        ```
        $ clana distribution --gt gt.csv --labels labels.csv [--out out/] [--long]
        ```
        
        prints one line per label, e.g.
        
        ```
        60% cat (56789 elements)
        20% dog (12345 elements)
         5% mouse (1337 elements)
         1% tux (314 elements)
        ```
        
        If `--out` is specified, it creates a horizontal bar chart. The first bar is
        the most common class, the second bar is the second most common class, ...
        
        It uses the short labels, except `--long` is added to the command.
        
        
        ### Metrics
        
        ```
        $ clana metrics --gt gt.csv --preds preds.csv
        ```
        
        gives the following metrics by
        
        * Line 1: Accuracy
        * Line 2: Precision
        * Line 3: Recall
        * Line 4: F1-Score
        * Line 5: Mean accuracy
        
        ### Visualizations
        
        See [visualizations](visualizations.md)
        
        
        ## Development
        
        Check tests with `tox`.
        
Keywords: Machine Learning,Data Science
Platform: Linux
Classifier: Development Status :: 4 - Beta
Classifier: Environment :: Console
Classifier: Intended Audience :: Developers
Classifier: Intended Audience :: Science/Research
Classifier: Intended Audience :: Information Technology
Classifier: License :: OSI Approved :: MIT License
Classifier: Natural Language :: English
Classifier: Programming Language :: Python :: 3.6
Classifier: Topic :: Scientific/Engineering :: Information Analysis
Classifier: Topic :: Software Development
Classifier: Topic :: Utilities
Description-Content-Type: text/markdown
