New Service Claims to Offer Deep Dive into Your Web Site’s Content
How Attivio Active Intelligence Engine’s next gen search works.
As publishing Web sites get more sophisticated, their true value won’t lie just in the breaking news stories at the top, but in the years of layered content and data. A new service, the Attivio Active Intelligence Engine, claims to offer publishers and their readers a way to perform a deep dive into a site—not just to find specific stories, but identify trends and data across many different pieces of content.
The service begins with the concept that there are two fundamental silos of data: numeric data (such as last month’s sales reports) and unstructured content, such as news articles and blogs.
“Enterprise search was fueled by explosive growth of unstructured content,” said chief technology officer Sid Probstein. “Our approach is to help companies with digital assets they want to monetize or make available for consumption.”
How it Works
The Active Intelligence Engine generate facets or dimensions that enable the reader to narrow their search while at the same time setting the basis for cross-links. “This is common in e-commerce,” Probstein said. “You may search for ‘digital camera’ but what you want is something with very specific specs. Facets help the user find their way to what they were looking for without requiring them to think about how to characterize it. With portal sites, facets are incredibly important. As search criteria, you don’t want to offer authors or dates but subjects and concepts.”
AIE uses an algorithm that helps define facets while taking the load off the editorial staff when it comes to tagging. The service also provides a spell-checking dictionary that can suggest correct or alternate spellings.
As a demonstration Attivio uses a search based on 100 years of baseball statistics. AIE was able to extract names of players from news stories and correlate those names with various statistics, ranging from salaries to how many home runs players in a certain weight class tend to hit. “This has the ability to recommend facets to be used as navigators to refine the search, such as people, authors, dates of article,” said Probstein.
Pricing and Requirements
The AIE runs on typical commodity hardware. One Java developer can quickly implement rich media functionality, Probstein said.
Pricing for the AIE can be either subscription-based or an upfront fee that scales up depending on functionality, he said. Pricing is typically “in the four-figure to six-figure range per month, depending on the applications used.”