Performance metrics
DataRobot offers a wide range of performance metrics for the models. You have to specify the metric you want to use to optimize the models for your project. Typically, the best metric to use is the one recommended by DataRobot. DataRobot does compute the other metrics as well once the model is built, so you can review the results of your model across multiple metrics. Please keep in mind that no metric is perfect for every situation, and you should be careful in selecting the metric for evaluating your results. Listed here are some details regarding commonly used metrics:
- RMSE (Root Mean Squared Error): RMSE is a metric that first computes the square of errors (the difference between actual and predicted). These are then averaged over the entire dataset and then we compute a square root of that average. Given that this metric is dependent on the scale of the values, its interpretation is dependent on the problem. You cannot compare RMSE for two different...
