Understanding the results
In this section, we will discuss various visualizations of metrics and other information to understand the results of the modeling exercise. These are important visualizations that need to be inspected carefully in addition to looking at the model metrics discussed in the previous section. These visualizations are generated automatically by DataRobot for any model that it trains.
Lift chart
The lift chart shows how effective the model is at predicting the target values. As the number of data points is typically very large to show in one graphic, the lift chart sorts the output and aggregates the data into multiple bins. It then compares the averages of predictions and actuals in each bin (Figure 2.13):
The preceding lift chart shows how the predictions have been sorted from low to high and then binned (60 bins in this case). You can now see the average prediction and average actual value in each...