Scikit-Learn Classifier Tuning from Simple Training Sets
Tuning is the process of maximizing an algorithm’s performance without overfitting, underfitting, or creating high variance. Overfitting is when an algorithm trains data so exactly that it may fail to fit new data or predict future results reliably. Overfitting usually occurs when a model is too complex for the data it is trying to train. An overly complex model trains data very well, but also fits noise that is not a part of the data. So, when used to train new data, noise is introduced causing unpredictable results.