.. _binary-performance-estimation:

================================================
Estimating Performance for Binary Classification
================================================

We currently support the following **standard** metrics for bianry classification performance estimation:

    * **roc_auc**
    * **f1**
    * **precision**
    * **average_precision**
    * **recall**
    * **specificity**
    * **accuracy**

For more information about estimating these metrics, refer to the :ref:`standard-metric-estimation` section.

We also support the following *complex* metrics for binary classification performance estimation:

    * **confusion_matrix:** a metric that has four components (TP, FP, FN, TN)
    * **business_value:** a metric that combines the four components of the confusion matrix using
      user-specified weights for each element, allowing for a connection between model performance and
      business results.

For more information about estimating these metrics, refer to the :ref:`confusion-matrix-estimation` and :ref:`business-value-estimation` sections.

.. toctree::
   :maxdepth: 2

   binary_performance_estimation/standard_metric_estimation
   binary_performance_estimation/confusion_matrix_estimation
   binary_performance_estimation/business_value_estimation
   binary_performance_estimation/custom_metric_estimation