Understanding H2O AutoML Leaderboard and Other Performance Metrics

When we train ML models, the statistical nuances of different algorithms often make it difficult to compare one model with another model that is trained using a different algorithm. From a professional standpoint, you will eventually need to select the right model to solve your ML problem. So, the question arises: how do you compare two different models solving the same ML problem and decide which one is better?

This is where model performance metrics come in. Model performance metrics are certain numerical metrics that give an accurate measurement of a model’s performance. The performance of a model can mean various things and can also be measured in several ways. The way we evaluate a model, whether it is a classification model or a regression model, only differs by the metrics that we use for that evaluation. You can measure how accurately the model classifies objects by measuring the number of correct...