Eqentia added to its content discovery and knowledge management portal this week features to recommend additional content or people connections to end users and content curators. But it’s also been doing some other interesting work in the past couple of months on the back-end that draws on semantic technologies to help curators and content administrators of custom Eqentia-based knowledge portals with their taxonomies.
This is where YAGO (Yet Another Great Ontology), a semantic knowledge base some 2 million entities strong that extracts structured information from Wikipedia via DBpedia, comes into play. In essence, YAGO reveals Wikipedia to the Semantic Web, explains CEO William Mougayar.
YAGO gives Eqentia a list of companies and persons to to use for its auto-complete list. Once the user clicks what he wants from the auto complete list — say “Steve Jobs”– Eqentia takes “Steve_Jobs” (note the underscore) and builds a SPARQL query to DBpedia that extracts all related labels by which DBpedia knows “Steve Jobs.” As Eqentia explains it, the upshot is that Eqentia uses a local copy of YAGO to quickly search companies and persons to get a unique “key” that is shared by all 3 systems (YAGO, DBpedia and Wikipedia), and which is then used to query DBpedia for any related labels.
“It’s a productivity benefit for the curator or content administrator when building custom vocabularies for the semantic tagging that we do,” he says. Eqentia, by the way, has close to 25 enterprises paying for Eqentia as the backbone for their own content portal infrastructures, Mougayar says. “The SPARQL query we use combines various ways a company or person can be labeled semantically, and we automatically create a synonym list that is assimilated by Eqentia.”
The capability removes a lot of the manual labor that would be otherwise required by the curator or administrator. In the works is using the GoodRelations ontology as another tool to help with taxonomy building.
Eqentia also has made it possible to move taxonomy vocabularies between various portals and topics, so that curators (as well as Eqentia itself) can more easily re-use existing taxonomies. A taxonomy editor lets them move an entity (such as ‘Subject’ in the screen above) from one portal to another, and edit it there, thus customizing the taxonomy to fit the new portal. As an example, a curator might want to leverage a taxonomy developed for cloud computing when she decides to create one for the related field of software-as-a-service (SaaS), speeding the process. “It sounds trivial, but it’s a great productivity saver. It’s like a copy/paste with edit function on taxonomies, vocabularies and synonyms,” Mougayar says.
The company also recently went live with its Drupal plug-in, following on the heels of its WordPress plug-in, which synchs up its taxonomy with Drupal’s taxonomy for sites that use the blogging platform and Eqentia’s curated articles.
The features introduced this week will be helpful both to curators and end users. “We are using key-phrase extraction with real-time semantic analysis to present content recommendations for the user or the curator,” says Mougayar. Users who’ve set up their personal news page with topics can hit the Recommended Content button to fetch 30 new articles from the last 24 hours that are related to their tracked topics. Another entry to content recommendations is synching a user’s Twitter account to her Eqentia personal stream, so that Recommended Content will be delivered based on any articles that she favorites on Twitter.
It’s also using the same Recommendations algorithm to suggest content for the curator of a topic, “so it’s an instant content boost for them…a productivity gainer,” he says.
On the people discovery front, Eqentia is in the social media mix now with capabilities that let users find key people engaging with their content on Twitter, Facebook and LinkedIn. It includes a Twitter Engagement Leaderboard listing the top 100 users that are interacting with a specific context, ranked by their engagement power or Klout score. It offers a Visual Who Is to display these users graphically and rank them according to a Klout-based influence score. It also provides insight into people’s implicit interest graphs through its aggregation of a user’s social presence in other interest areas.
Says Mougayar, “We want to continue to push the envelope in terms of enterprise-level strength and content curation.”