Entropy_triangle is a function that implements two tools to analyze the behavior of multiple-class, or multi-class, classifiers by means of entropic measures on their confusion matrix or contingency table. First we obtain a balance equation on the entropies that captures interesting properties of the classifier. Second, by normalizing this balance equation we first obtain a 2-simplex in a three-dimensional entropy space and then the de Finetti entropy diagram or entropy triangle.
A full description can be found in:
Valverde-Albacete, F. and Peláez-Moreno, C. Two information-theoretic tools to assess the performance of multi-class classifiers. Pattern Recognition Letters (2010) vol. 31 (12) pp. 1665-1671
In version 2.0 the entropy triangle is added a color bar to plot another variable, e.g. accuracy, against mutual information, variation of information of entropy decrement. We have also added a script (compareETs) to visualize your own ETs and to print the NIT and EMA vs. other measures in latex-ready format.
The Normalized Information Transfer (NIT) rate, and the Entropy-Modified Accuracy (EMA) are improvements to accuracy to measure how good your multiclass classifiers are whose rationale and definition can be found in:
F. J. Valverde-Albacete and C. Peláez-Moreno. 100% classification accuracy considered harmful: the normalized information transfer factor explains the accuracy paradox. PLOS ONE, 9(1):e84217, January,
Francisco José Valverde-Albacete (2022). Entropy triangle (https://www.mathworks.com/matlabcentral/fileexchange/30914-entropy-triangle), MATLAB Central File Exchange. Recuperado .
Compatibilidad con la versión de MATLAB
Compatibilidad con las plataformasWindows macOS Linux
Inspirado por: alchemyst/ternplot
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!