Gurucul is providing details on a few of our most popular machine learning models. Check out what’s up next!
Gurucul Machine Learning Model: Dimensionality Reduction
How does the Dimensionality Reduction machine learning model work, what does it do? The Dimensionality Reduction machine learning model performs line and attribute filtering. It reduces raw logs. By doing this work for you, Gurucul Risk Analytics gets rid of 80% of what is essentially useless data.
Gurucul Risk Analytics feeds data into the Dimensionality Reduction machine learning model to determine which attributes should be tracked and analyzed. You don’t need to monitor all the data all the time. You only need to examine data that will give you measurable results when analyzed by our security analytics engine.
Use Case: Anti-Money Laundering
Anti-money laundering is a typical fraud use case. Gurucul Risk Analytics tracks transactions for money transfer and will determine if someone in your organization is making a large number of transactions of a small dollar volume. If there are repeated patterns, for example $999 every day to the same place or same account. Or, if we see the money is being sent to risky locations like Iraq, Afghanistan or Syria, this is more than likely money laundering activity.
The interesting thing about financial transaction data is that it contains a lot of characteristics that are not relevant for analytics. Out of 50 attributes, usually only 5% may be qualified for analytics. Gurucul Risk Analytics reduces the data source pool down using Dimension Reduction to a manageable essential number for use. There is no point in feeding or wasting time analyzing meaningless or bad data.
What are the Benefits of Dimensionality Reduction?
Dimensionality Reduction reduces the amount of data stored and analyzed. This is a very big deal. You don’t want to store or spend time wading through useless data.
In terms of anti-money laundering, Gurucul has been able to increase our detection rate. In fact, in a recent POC, Gurucul Risk Analytics detected more financial fraud than IBM Watson. Ultimately, this will results in saving millions of dollars for your organization.