Weighted one-against-all
Alina Beygelzimer, John C. Langford, Bianca Zadrozny · 2005
The one-against-all reduction from multiclass classifi-cation to binary classification is a standard technique used to solve multiclass problems with binary classi-fiers. We show that modifying this technique in order to optimize its error transformation properties results in a superior technique, both experimentally and the-oretically. This algorithm can also be used to solve a more general classification problem “multi-label classi-fication, ” which is the same as multiclass classification except that it allows multiple correct labels for a given example.