From 25dc4730e9ba68b561888d99ec68ce74282f3aa0 Mon Sep 17 00:00:00 2001 From: Fabian Kueppers Date: Thu, 28 Nov 2024 16:48:59 +0100 Subject: [PATCH] Minor visualization improvement in performance safety aspects --- docs/source/evaluation-aspects/performance.rst | 1 + 1 file changed, 1 insertion(+) diff --git a/docs/source/evaluation-aspects/performance.rst b/docs/source/evaluation-aspects/performance.rst index 54d75cf..b8dd2ba 100644 --- a/docs/source/evaluation-aspects/performance.rst +++ b/docs/source/evaluation-aspects/performance.rst @@ -27,6 +27,7 @@ Currently, the following metrics are implemented: * **Matthews correlation coefficient (MCC)** is a metric used to evaluate the performance where different aspects are gathered within a single score. The metric ranges from -1 to +1, where a score of +1 indicates perfect predictions with no mistakes. A score of 0 suggests that the classifier's predictions are no better than random guessing, while a score of -1 indicates total disagreement between predictions and actual labels, with all predictions being incorrect. In the context of binary classification, we also report the following metrics: + * **Specificity** complements sensitivity by measuring the classifier's ability to avoid false alarms in identifying negative instances. It quantifies the proportion of true negative instances that are correctly identified out of all actual negative instances. High specificity minimizes false positive errors, enhancing the classifier's ability to accurately identify negative instances. * **Negative predictive value** assesses the reliability of negative predictions made by the classifier. It quantifies the proportion of true negative predictions out of all negative predictions made by the classifier. Negative predictive value is particularly relevant in scenarios where correctly identifying negative instances is crucial. * **Informedness** measures the classifier's ability to make correct positive and negative predictions simultaneously. Informedness ranges from -1 to 1, where a score of 1 indicates perfect classification, 0 indicates random classification, and -1 indicates perfectly incorrect classification.