In Part I, we discussed how the cross-entroy cost function was a more robust method to compare the effectiveness of classification models by way of example. Toward the end of that discussion, we presented an equation and showed how it could be used to do a better job of comparing models that using the overall accuracy.
In this part, we’ll introduce the idea of maximum likelihood which is a general approach to building cost functions that are used to estimate model parameters.
To be continued…