Refinder and Stanbol Integration Showcase Launched

Today, we are very happy to announce the launch of the Refinder/Stanbol integration showcase.

The purpose of the Refinder/Stanbol integration is to showcase the usage of Apache Stanbolsoftware components to add automatic tagging functionality to both Refinder services as described before. In this context, by “tags” we denote resources, which are identified using HTTP URLs, and feature rich metadata (including multilingual labels and relations to other resources) expressed in RDF. Tags are generated by Stanbol’s enhancer service, which takes as input the full textual description of items in Refinder (including possibly related content and user-generated content like comments). Stanbol analyzes the content and matches it to resources from DBpedia, which are in turn represented as tags in Refinder. These tags can be used to find and retrieve content, whereas (through the usage of tag descriptions) more diverse search options are available to the user. For instance, it is now possible to search for tags in multiple languages, since translations for tag labels are provided by DBpedia.

Demos are available at Refinder for Teams and Refinder Cloud Search. A demo account (which has already been populated with data from different sources, including RSS feeds, Twitter streams, and Dropbox files) exists:

Currently, the Stanbol-based enrichment service is not activated by default for every user, so if you want to try it with your own Refinder account, contact us at so that we can enable the enrichment service for you.

To see the service “in action”, issue a search request, e.g., for “social media“. In the result list, you will see tags that are associated to the result documents. These tags are clickable: by clicking on a tag the results will be reduced to items that are associated with that tag. (The same functionality can be triggered by selecting a tag from the “Tag” filter panel on the left side of the screen.

In this demo, we can see that the usage of semantic annotations for information items (in this case, RSS feeds, Twitter streams, and Dropbox files) greatly reduces the time needed for search and retrieval. Additionally, it increases the search precision since it allows to search using synonyms, acronyms, and alternative languages, which are mapped to semantic tags that are associated to the indexed information items.

There is some work left to do regarding the performance of semantic tag descriptions on the Refinder side. As soon as they are solved, the Stanbol tagger will be released to all Refinder users. Stay tuned!