IBM enhances analytic capabilities within i2 Safer Planet Intelligence Analysis portfolio

The IBM portfolio lets users search across myriad data sources for information and identifies links between people and assets—for instance, a vehicle that was used in a drug-trafficking operation—that probably would not be evident on the surface to investigators, Dalzell said.

In addition, the visual depiction in the network diagram can help a user determine the likelihood of a link, so resources can be utilized in the most efficient manner.

As a fun example, Dalzell used the IBM algorithm engine to search the IMDB database to determine whether Kevin Bacon knows Clint Eastwood and Tom Cruise. While there was a direct link between Bacon and Cruise—both actors starred in the film, A Few Good Men—a connection between Bacon and Eastwood was much more convoluted. However, all of the potential paths in the IMDB database between the two men had a common link, Morgan Freeman.

“Imagine that we’re looking at a terrorist network,” Dalzell said. “What I can tell right away is that these two people—Kevin Bacon and Tom Cruise—are linked by this one event here, so the hypothesis I can draw is that, with 99% certainty, they know each other. And that’s interesting, because now I don’t necessarily need to focus any more investigation on that part of the story.

“But, if I look over here, Clint Eastwood is connected to [Bacon via] multiple people, and Morgan Freeman is the gatekeeper between these two ranges of films. Now, I can’t say with any certainty that [Bacon and Cruise know Eastwood], so I would need to carry out further investigation in this part of the network to understand. Now, I can focus my effort on this part of the network—and I can conclude that … [Morgan Freeman] would be the best place for me to start the investigation.”

For high-profile people like those used in the example, an analytics tool probably is not needed to deduce such connections. However, the ability to let computer-generated algorithms quickly find connections between people that do not want to be found is invaluable, Dalzell said.

“It’s very cool. I used to have to do this manually … and it took weeks to work this out,” he said. “Previously, you would have had to have done a series of manual find paths through the data to find this stuff. Using the mathematics, this was hard. Now it can be done in seconds—it takes four [mouse] clicks—and it’s formatted, so you can see [the link] really quickly.”