Rare and Volume Based Analytics

Prevent Fraud with “Rare and Volume Based Analytics”

We are pleased to provide details on our most popular machine learning models. Check out the next in our series of informative blog articles.

Gurucul Machine Learning Model: Rare and Volume Based Analytics

How does the Rare and Volume Based Analytics machine learning model work, what does it do?  This machine learning model identifies rare activity based on volume.

Across most businesses, it’s normal for a customer support representative to carry a support load of about 20 customers. Let’s say – all of a sudden, your customer support representative is now looking at 50 different customer records. This abrupt increase in volume of access activity is unusual. The customer support representative may not be authorized to manage, and thus access, this volume of customer records. Gurucul Risk Analytics would flag these activities as suspicious via this rare and volume based behavior.

Use Case: Fraud Detection & Prevention

Rare and Volume Based Analytics is one of many machine learning models included in Gurucul Risk Analytics that is used to detect and prevent fraud. At Gurucul, we like to call it our “snoop finder”, as it roots out and exposes those uncomfortable situations where users on your network may be using their own account, or someone else’s, to snoop around and look at or gather information present inside your network.

Let’s look an example. Consider a medical office setting where a particular doctor has not typically accessed patient records. All of a sudden, the doctor, or his user account, begins accessing patient records. In fact, his user account appears to be crawling through several patient records. That’s anomalous activity.

This same set of behavior analytics applies to customer support fraud as it’s the same model at work behind the scenes. Typically, it’s only when a customer calls into a support center that a customer support representative views a customer’s record. So, if a customer support rep begins looking at customer records without an apparent purpose that may reveal fraudulent behavior.

What are the Benefits of Rare and Volume Based Analytics?

There are several tangible benefits to having the ability to detect and prevent fraudulent behavior in your organization before it happens.  Beyond the obvious pain and costs of having monies inappropriately transferred or records being stolen, there are tremendous organizational costs to root out the perpetrators, to perform any recoveries, and to regain the confidence of your customers.  Some relationships may not be recoverable.

Gurucul Risk Analytics is able to detect and prevent fraud before money is transferred or records are stolen through the Rare and Volume Based Analytics machine learning model.  This powerful model can also be used to satisfy regulatory and compliance mandates, such as HIPAA and PCI-DSS.

We invite you to watch this video to understand how Allina Health used this model (and others) to protect both Players’ and VIP medical records during the 2018 Super Bowl in Minneapolis. Gurucul’s Risk Analytics was able to answer these questions: “Are you snooping on other employees? Are you looking at VIP records?” Gurucul’s behavior based security analytics models link separate events together and triggers their coincidence as an alert to say: “We have an event where a VIP record is being accessed in combination with data being exfiltrated from that VIP record”. These alerts will point you to potentially troubling activities that you seriously need to take a look at right away.

Share this page:
Previous
Next