In-Depth

With Demand Red-Hot for Text Analytics, Attensity Expands Suite

Demand for text analytics has been red-hot, fueled by compliance, fraud detection, and other emerging drivers.

• By Stephen Swoyer
• 10/11/2006

If black is the new blonde, then text mining is—in a similar sense—the latest and greatest of killer apps. This is probably due in part to a broader trend in favor of data mining, especially now that more and more non-specialty vendors have incorporated data mining or other analytic capabilities into their relational database platforms and bread-and-butter business intelligence (BI) applications.

But demand for text analytics, in particular, has been red-hot, fueled by compliance, fraud detection, and other emerging drivers.

Enter text analytics specialist Attensity Corp., which recently shipped a new version of its all-in-one text mining suite, Attensity 4, that includes enhanced routines for mining, searching, querying, charting and graphing unstructured text, along with several new application modules.

In putting together the revamped Attensity suite, officials say they’ve taken customer suggestions to heart. "We consulted extensively with our government corporate customers to create the first Text Analytics suite that enables seamless decision support on your text without requiring extensive knowledge engineering. Attensity 4 puts the full power of Text Analytics into the hands of the business user," said Attensity chief Craig Norris, in a prepared release.

It’s an important deliverable for Attensity, says James Kobielus, a principal analyst for data management with consultancy Current Analysis. First, he says, it gives Attensity "a comprehensive text analytics suite that includes new methods for mining, searching, querying, charting and graphing free-form unstructured text dynamically from within a browser-based interface." More importantly, Kobielus notes, the revamped Attensity 4 consolidates the company’s existing (and to some degree desultory) product line on a unified architecture and data model.

What’s more, Attensity 4 updates existing Attensity’s Discover and Analytics modules, while at the same time delivering new modules: in this case, Attensity Text Search, Attensity Classify, Attensity Manage and Attensity Integration. All of Attensity’s modules, new and existing alike, can be licensed either together or separately, Kobielus points out.

Nor is that all: Attensity has also enhanced its server-side fact-extraction engine by combining a patented new technique, dubbed Exhaustive Extraction, with other targeted engines, such as Entity Extraction, Event Extraction, Keywords, and Directed Learning extraction. Finally, Kobielus notes, there’s the requisite SOA tip: customers can expose Attensity 4’s text-analytics services via well-defined interfaces and embed text analytic capabilities into either packaged or custom applications.

For these and other reasons, Kobielus argues, Attensity 4 is an important deliverable—both for Attensity itself and for the burgeoning text analytics marketplace in general. "Attensity’s announcement was a necessity for the vendor to continue distinguishing itself in the crowded, growing, and very competitive text-analytics segment. Attensity now provides a comprehensive text analytics suite, incorporating, enhancing, and expanding upon the products that it had previously provided as stand-alone tools," he argues, noting for example that the revamped suite includes a text-search feature which Attensity has licensed from a third-party vendor.

In addition, Kobielus points out, Attensity 4 "provides all functional components within a unified architecture, data model, processing workflow model, role-based security, administrative toolset, alerting services, and browser-based user interface." Finally, Attensity’s revamped Discover module (which provides fact-extraction and analysis capabilities) now supports linguistics-based algorithmic approaches, too.

Just as important, Kobielus notes, Attensity has ratcheted up its partnering efforts, co-announcing (in tandem with the Attensity 4 release announcement) partnerships or joint marketing agreements with Business Objects SA, IBM Corp., and Teradata (a division of NCR Corp.).

So much for the good. The bad—or, rather, the not so good—is that few if any of the best-of-breed text analytic vendors have yet established their offerings as viable alternatives to competitive offerings, such as a range of products now marketed by enterprise content management (ECM) specialists.

"Attensity and other text analytics vendors have not yet effectively positioned their technology as an alternative to… ECM platforms, knowledge management products, and search engines, which also enable users to quickly extract keywords and other entities from unstructured text," he points out.

And the bigger rub, at least from Attensity’s perspective, is that a number of BI power-players—including SAS Institute Inc., SPSS Inc., Oracle Corp., and IBM Corp.—already provide text analytic capabilities as part of their BI, predictive analytics, or data mining portfolios. While Attensity and other best-of-breed competitors (such as Clarabridge, Clearforest, Factiva, and Inxight) can still credibly claim to deliver capabilities that outstrip those available from most of the mainstream players, that might be a moot point, Kobielus suggests.

"Larger… vendors of BI, predictive analytics, data mining, ECM and other products should consider acquiring text-analytics vendors—perhaps Attensity—to round out or enhance their product suites," he concludes.