J. E. Chacón
The adjusted Rand index (ARI) is a measure of accuracy for clustering evaluation that is very commonly used. Many of the users, however, seem not to be fully aware of its meaning, and this frequently leads to misinterpretations that invalidate the conclusions thus obtained from simulation or experimental studies. In this talk, we will review closely some properties of this index, and will compare it with suitable alternatives.
Palabras clave: clustering evaluation, adjusted Rand index, misclassification error
Programado
GT18 Estadística no paramétrica II. Inferencia no paramétrica para la densidad y la distribución
9 de junio de 2022 12:00
A15
