Diferencias

Muestra las diferencias entre dos versiones de la página.

--- clase:iabd:pia:2eval:tema07.metricas_derivadas [2024/03/25 14:33]
admin [Juntado dos Métricas Básicas]
+++ clase:iabd:pia:2eval:tema07.metricas_derivadas [2024/03/25 14:47] (actual)
admin [Juntado dos Métricas Básicas]
@@ Línea 175: / Línea 175: @@
   * [[https://machinelearningmastery.com/roc-curves-and-precision-recall-curves-for-classification-in-python/|How to Use ROC Curves and Precision-Recall Curves for Classification in Python]]
   * {{ :clase:iabd:pia:2eval:predicting_receiver_operating_characteristic_curve_area_under_curve_and_arithmetic_means_of_accuracies_based_on_the_distribution_of_data_samples.pdf |Predicting Receiver Operating Characteristic curve, area under curve , and arithmetic means of accuracies based on the distribution of data samples}}
-  * [[https://neptune.ai/blog/f1-score-accuracy-roc-auc-pr-auc|F1 Score vs ROC AUC vs Accuracy vs PR AUC: Which Evaluation Metric Should You Choose?]]
+  * Cálculo del mejor Threshold:
-  * [[https://stackoverflow.com/questions/44172162/f1-score-vs-roc-auc|F1 Score vs ROC AUC]]
+    * [[https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5470053/|Defining an Optimal Cut-Point Value in ROC Analysis: An Alternative Approach]]
+    * [[https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5082211/|On determining the most appropriate test cut-off value: the case of tests with continuous results]]
+<note tip>
+Hay otra curva que en vez de ser (1-Especificidad) vs Sensibilidad , es la de Sensibilidad vs Precisión (llamada en inglés Precision-Recall) que se usa cuando los datos tienen una baja prevalencia.
+Y además está relacionado con el F1-score ya que el F1-score se calcula justamente con la Sensibilidad y Precisión
+  * [[https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4349800/|The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets]]
+  * {{ :clase:iabd:pia:2eval:roc_graphs_notes_and_practical_considerations_for_researchers.pdf |ROC Graphs:Notes and Practical Considerations for Researchers}}
+  * [[https://juandelacalle.medium.com/how-and-why-i-switched-from-the-roc-curve-to-the-precision-recall-curve-to-analyze-my-imbalanced-6171da91c6b8|How and Why I Switched from the ROC Curve to the Precision-Recall Curve to Analyze My Imbalanced Models: A Deep Dive]]
+  * [[https://en.wikipedia.org/wiki/Mann%E2%80%93Whitney_U_test#Area-under-curve_(AUC)_statistic_for_ROC_curves|Area-under-curve (AUC) statistic for ROC curves]]
+  * F1-score y ROC
+    * [[https://neptune.ai/blog/f1-score-accuracy-roc-auc-pr-auc|F1 Score vs ROC AUC vs Accuracy vs PR AUC: Which Evaluation Metric Should You Choose?]]
+    * [[https://stackoverflow.com/questions/44172162/f1-score-vs-roc-auc|F1 Score vs ROC AUC]]
+</note>
 ===== Juntado dos Métricas derivadas =====
 Las 4 métricas derivadas son PPV, NPV, FDR y FOR.

logongas

Herramientas de usuario

Herramientas del sitio

Diferencias

Herramientas de la página