Tools written in R for doing security

Cybersecurity data analysis

Model evaluation

Run this code to:

1. Create a basic random forest classification model.
2. Fit that model to the latest VERIS Community Database data.
3. Test the model's predictive accuracy on using more historical data.
4. 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)

VCDB-Partial-Dependence-Plots.R

More

• http://veriscommunity.net/discovery.html#section-incident-timeline
• https://stephenmilborrow.r-universe.dev/plotmo