2013 R&D 100 Winner
Distribute the Highest Selected Textual Recommendation (DTHSTR)Every person, company and government organization is faced with continuously flowing, massive streams of data that can’t be manually analyzed. Computers help, but the sheer volume of data has foiled many approaches. Software engineers at Oak Ridge National Laboratory have now developed a new way of dealing with this seemingly intractable problem.

The Distribute the Highest Selected Textual Recommendation (DTHSTR) engine combines advanced analytic technology with a low computing footprint, allowing for analysis of large and dynamic text data with unprecedented speed and accuracy. The engine is designed to help users find documents related to their interest regardless of the application domain. DTHSTR operates by leveraging example documents or information provided by a user as a means of defining what content is relevant to the user and creates a profile based on this content. Next, DTHSTR analyzes this profile to identify significant words and phrases that are then used to perform a series of searches in a much larger target data set in order to find other documents with these significant words and phrases. Documents returned from the series of searches are then compared to the user profile in order to identify the most relevant documents and sort the result set according to similarity to the user profile.

Search engine

Oak Ridge National Laboratory

Development Team

Oak Ridge National Laboratory's Distribute the Highest Selected Textual Recommendation (DTHSTR) development team (l-r): Chad Steed, Thomas Potok, Robert Patton.


The Distribute the Highest Selected Textual Recommendation (DTHSTR) Development Team from Oak Ridge National Laboratory
Robert Patton, Principal Developer
Thomas Potok
Chad Steed
Jim Treadwell