For the error plot, as this is an online method, I consider every sample implicitly as a test one. The goal was to show that after N correction, the algorithm c

For the error plot, as this is an online method, I consider every sample implicitly as a test one. The goal was to show that after N correction, the algorithm converges to a weight configuration that is compatible with a specific data distribution. As I wrote, this can be a limitation (common to many online methods), because it's almost impossible to filter outliers out and the tolerance to noise is surely worse than an offline algorithm.