By Ross Mauri
General Manager, IBM z Systems
Remember those college days of cramming for final exams, frantically trying to recall every piece of information on a topic. Imagine instead walking into the classroom with the world’s top expert by your side and the ability to tap his or her knowledge at any moment in time.
The chances of this happening in college are far-fetched, but in many ways the business world is about to undergo this type of transformation. The emerging field of machine learning has the potential to process data faster than humanly possible. And when connected with the right sources of information, the combination creates a constantly learning subject-matter-expert – a doctoral candidate at Harvard, if you will.
As a major step toward making this new paradigm a reality, IBM announced today that it is pairing together two advanced technologies – Watson’s machine learning and the IBM z Systems mainframe. IBM Machine Learning for z/OS uses core machine learning technology from Watson to help organizations develop and improve analytic models based on their mainframe data.
Combining machine learning with the mainframe is significant because the z Systems mainframe holds some of the most valuable data for major organizations around the world. The mainframe is the core processing system for 44 of the top 50 global banks, ten of the top insurers and 18 of the top 25 retailers. Think of machine learning as the eager student wanting to learn more and the mainframe as the library filled with every book and report on the subject.
Analyzing and applying all of the data on the mainframe would be challenging for even the most skilled data scientist, but machine learning can help with the heavy lifting. Machine learning can process data on a massive scale, sorting through millions of pieces of data on the mainframe to surface connections and present scenarios that would be difficult to find otherwise.
What’s more, machine learning uses systems that can improve over time through continuously updated data. It not only surfaces more insights but constantly seeks to improve the accuracy of the analytic models. IBM Machine Learning for z/OS also supports popular machine learning frameworks and programing languages to make the solution widely accessible.
So how can machine learning on the mainframe help a business? Let’s say, for example, you’re an outdoor retailer that uses the mainframe to process all transactions. You know certain products, like tents and insect repellent, sell well together, but the company has thousands of products. Preferences also vary greatly between locations and customers.
Machine learning can dig deeper by processing more data to find new connections. The business can move from a basic cross-selling of common products to tailored offerings for each customer. Because machine learning is run on the mainframe, the analysis occurs quickly, providing real-time updates to adjust inventory or incentives. Security standards are also maintained because data never leaves the mainframe.
The outdoor retailer is just one example of new possibilities from machine learning and the mainframe. It can also help health care providers personalize patient care or help financial advisors tailor investment choices.
Machine learning on the mainframe has the potential to create a new way of operating through a deeper understanding of markets and customer needs. Business leaders still have to use the new insights to take actions and guide their organizations, but it never hurts to walk into the office each day with a budding genius by your side.
For more information about IBM Machine Learning for z/OS and how it can help your organization, watch this video on the new offering.
(Link for embedding: https://youtu.be/T2HtyNX7aHc).
You also can check out new analytics capabilities through the IBM z Analytics Trial.