Exploring other model performance metrics
The H2O AutoML leaderboard summarizes the model performances based on certain commonly used important metrics. However, there are still plenty of performance metrics in the field of ML that describe different skills of the ML model. These skills can often be the deciding factor in what works best for your given ML problem and hence, it is important that we are aware of how we can use these different metrics. H2O also provides us with these metrics values by computing them once training is finished and storing them as the model’s metadata. You can easily access them using built-in functions.
In the following subsections, we shall explore some of the other important model performance metrics, starting with F1.
Understanding the F1 score performance metric
Precision and recall, despite being very good metrics to measure a classification model’s performance, have a trade-off. Precision and recall cannot both have high values...
