Predictive analytics is a subject we’re keeping an eye on this year. Today we’re taking a look at some of the ways predictive analytics are being used by businesses: to predict customer attrition rates, monitor IT systems and issue alerts when appropriate, optimize prices for auctions and prepare for GMAT tests.
Customer Attrition Rates
Last year Theresa Doyon wrote about how survival analysis could be used to predict customer attrition rates. Today, 1to1 Media posted a talk, embedded above, by Colin Shearer, worldwide industry solutions leader for the SPSS brand at IBM, on that very subject.
IT Systems Monitoring and Alerting
Last month we covered how Netuitive is using predictice analytics to make IT’s life easier. Netuitive monitors tens of thousands of metrics to make predictions about IT system failures and send intelligent alerts. Today Netuitive announced that has landed another major banking customer, but can’t reveal who that customer is. Netuitive claims to now provide its analytics solution to eight of the top ten largest banks.
Opera Solutions offers analytics software and consulting for a wide variety of problems. One example is auction pricing optimization built for a major automotive customer. Opera emphasizes interaction and human judgement. Instead of just aggregating and crunching a bunch of pricing information and giving auctioneers a recommended price, the application lets users input information from the auction floor to update pricing in real-time. In order to deal with unforeseen circumstances, Opera allows users to change the weight of different factors on the fly without being versed in statistics or programming.
Opera, as part of a group called The Ensemble, submitted an entry for the Netflix Prize that performed as well as the winning entry, BellKor’s Pragmatic Chaos. However, it was submitted 20 later later than Pragmatic Chaos so The Ensemble did not win.
The company has built a set of adaptive learning tools into its products similar to the recommendation engines of companies like Amazon.com and Netflix. Knewton’s software analyzes students’ performance on practice questions and recommends tutorials based on the student’s answers. Knewton optimizes learning by focusing only on the areas that students need to improve. The software determines subject areas at a granular level. it doesn’t just know whether you need improvement in algebra. It knows specifically whether you’re having trouble with, for example, quadratic equations.
According to COO David Liu, an afternoon of studying can give Knewton 100,000 – 150,000 data points about the student – such as how long it takes them to answer questions or what time of day they learn best.
Knewton is focused on the education market right now, but it plans on releasing an API this summer that would open its platform to more applications.
Image by loop_oh
Disclosure: IBM is a sponsor of ReadWriteWeb