Semantic Search
-
Case Study
Newssift's search engine exposes categories, people, topics, organizations and places, allowing a unique and intuitive browsing experience.
Powering search, semantically
Search isn’t just about presenting the single piece of relevant content (when it exists!) a user is looking for; people want answers to questions, rather than search results. Semantic Search is about answering questions. By indexing content and applying a rich layer of structured metadata to it, Nstein (coupled with a search engine) opens opportunities to create revolutionary new search experiences.
Nstein labels each piece of content according to categories, people, organizations, concepts, locations, even tone. A "linguistic fingerprint" of every piece of content is created, so that similar pieces of content can be automatically related. Users can then delve into those newly created relationships and facets (or semantic annotations) for navigation through search results to refine the document set until they find the information they are looking for.
FT Search’s Newssift uses Nstein’s TME product to power it search engine to mine content daily from over 4,000 sources of financial news - more than 70,000 pieces of content a day! Leveraging the semantic metadata generated by TME through its search engine makes for a highly relevant and accurate “finding” experience. Searchers are returned not only results to their search query, but also related items of interest organized by category, ensuring users get to the content they want, fast!