Predictive Analytics Now Easier – but Still No Cake Walk: Page 2
Updated · Mar 08, 2022
“To be successful, practitioners must be able to work effectively with the business side to ask the right questions and have a solid understanding of the limitation of predictive analytics techniques,” said Stefan Schmitz, vice president of Product Management at MicroStrategy, a provider of analytics software.
Predictive Analytics and Rapid Iteration
“Organizations should focus on a flexible analytics infrastructure that is not only able to meet a number of diverse requirements across the enterprise, but also one nimble enough to adapt quickly to the needs of the business,” Hillion said.
The agile software development movement can serve as a useful example. Instead of aiming for an ideal solution, focus on rapid development of fresh predictive analytics models that can be pushed quickly into production. Then monitor performance with testing and user feedback loops to continually test, refine and improve analytic applications and begin the cycle again.
There is a growing consensus that applying predictive analytics to huge, generalized issues is unwise. Imagine employing IBM Watson, for example, to find the secret of the universe. All that would happen would be that it would crunch data endlessly.
Many companies find that predictive analytics is most useful when applied to specific problems instead. A growing use case, for example, is detecting security breaches. These days, it is safe to assume that the bad guys have already breached the network. Predictive algorithms can “learn” the behavior patterns employed by criminal infiltrators in order to detect their presence.
“Analytics will drive better customer experiences by putting security safeguards in the background,” said Otto Berkes, CTO at CA Technologies. “New analytics techniques will use behavior patterns and machine learning to help separate the real customers from the fraudsters. Analytics isn’t ever going to be solution to anything, but increasingly, it’s going to be part of the solution to everything.”
Drew Robb is a freelance writer specializing in technology and engineering. Currently living in Florida, he is originally from Scotland, where he received a degree in geology and geography from the University of Strathclyde. He is the author of Server Disk Management in a Windows Environment (CRC Press).