Covid-19 and the Data
There is a lot of misinformation about Covid-19. I am fortunate to work with, what I consider, the best data modeling tool presently available.
Interset, is the name of the technology, which is being used to help provide a predictive model for Covid-19.
What is Interset? How do I use it for cyber security and why is it used for Covid-19 forecast models?
Interset is a user entity and behavior analytics UEBA, engine, and is able to ingest many types of data. For example, to be able to find a compromised account, I would be interested in collecting data from user logs, network flows, organizational folders, and applications. Separately, data from each entity listed, would not help uncover malicious intent in an organization. But a very powerful picture could be created if all the pieces were brought together in one model, and this is what Interset is able to complete. A baseline can be created for every entity, making a “unique normal”.
How is this helpful in cybersecurity? Rules and thresholds are created manually to establish alerts. When an alert has come into the queue, a person will open the alert and further investigate the incident. But often we have created rules that do not scale, causing too many false positives and alert fatigue. It would be similar to having a paint by numbers canvas and using a large paint brush, with only one color of paint. The results would lack detail, making it undistinguishable.
Interset uses multiple machine learning algorithms, as well as, unsupervised machine learning, to extract the information and provide in depth details about possible fraud, malicious insider threat, and even espionage. Supervised learning requires data that has already been labeled. This is easily accessed since the cyber community has decade’s worth of malware examples, both malicious and accidental. Unsupervised learning is a self-discovery of patterns.
So how does this help us provide a forecast for Covid-19? First, Interset is able to observe and quantify, establishing a “unique normal”. Second, Interset has been provided data; including death data from the Center for Systems Science and Engineering at John Hopkins University, and intervention data from Oxford Covid-19 Government Response Tracker, and Covid Local Analysis and Mapping of Policies, to build an application.
The application provides a visual model to help illustrate past, present and a fourteen day future prediction. Thus far, the model has been accurate and the data scientist at Interset hope that people can see what impact their actions may have for not only today, but for the future. Stay safely Connected.