Metadata-Version: 1.0
Name: confumatrix
Version: 0.1
Summary: Confusion matrix tool for binary classification.
Home-page: https://github.com/moliware/confumatrix
Author: Miguel Olivares and F. Javier Alba
Author-email: miguel@moliware.com
License: BSD
Description: confumatrix
        ===========
        
        Confusion matrix tool.
        
        Further information:
        
            * http://en.wikipedia.org/wiki/Binary_classification
            * http://en.wikipedia.org/wiki/Sensitivity_and_specificity
        
        Instalation
        -----------
        
        From source code: ::
        
          python setup.py install
        
        From pypi: ::
        
          pip intall confumatrix
        
        
        Usage
        -----
        ::
        
            >>> from confumatrix import ConfuMatrix
            >>> cm = ConfuMatrix(a=10, b=1, c=10, d=3)
            >>> cm.accuracy()
            0.5416666666666666
            >>> cm.precision()
            0.9090909090909091
            >>> cm.recall()
            0.5
            >>> cm.specificity()
            0.75
            >>> cm.sensitivity()
            0.5
            >>> cm.f1()
            0.6451612903225806
            >>> print cm
            =========================
            Confusion Matrix:
        
                10  1
                10  3
        
            Accuracy: 0.541666666667
            Recall: 0.5
            Precision: 0.909090909091
            Specificity: 0.75
            F-measure: 0.645161290323
            =========================
        
        
        binary_classification.py
        ------------------------
        
        Load a yaml file with results of a binary classfication and prints Confusion matrix for that results.
        
        Input file format: ::
        
            # Languages that can be detected
            types: [en, es, pt, it, de]
            umbral: 0.5
            sample: 
              - result:
                  en: 0.1
                  es: 0.2
                  pt: 0.3
                  it: 0.6666  #hit! ( 0.6666 > 0.5 )
                  de: 0.3
                expected: de
              - result:
                  en: 0.2
                  es: 0.99 # hit
                  pt: 0.98 # hit
                  it: 0.4
                  de: 0.3
                expected: es
Platform: UNKNOWN
Classifier: Development Status :: 4 - Beta
Classifier: Natural Language :: English
Classifier: License :: OSI Approved :: BSD License
Classifier: Programming Language :: Python :: 2.5
Classifier: Programming Language :: Python :: 2.6
Classifier: Programming Language :: Python :: 2.7
