Run this code to:
- Create a basic random forest classification model.
- Fit that model to the latest VERIS Community Database data.
- Test the model's predictive accuracy on using more historical data.
- Visualize how Action types predicted Time-to-Discovery.
We'll visualize the model's predictive ability by plotting 7 Partial Dependence Plots (PDPs), one for each of the categories of "Action" in the VERIS schema. Our predictors: malware, hacking, social, misuse, physical, error, environmental Our predicted: Time to Discovery (timeline.discovery.value)
