Semaphore for Microsoft FAST ESP - Better Search Navigation & Content Classification
Semaphore Enterprise Semantic Platform Delivers Semantic Search
Semaphore has been integrated with a number of enterprise search applications, including the Google Search Appliance (GSA), Solr, and Microsoft FAST and others, to add semantic processing capabilities to these search engines to provide:
- Accurate and consistently applied metadata and classification.
- An ontology-driven solution that improves findability.
- A better search-navigation experience, whatever the search engine.
Benefits of Microsoft FAST Solution
Improved Search Navigation and Classification Capabilities with Semaphore for FAST ESP
Semaphore for FAST ESP delivers an enhanced search experience to users of the FAST Enterprise Search Platform, with model-based navigation and improved search results.
Semaphore Enterprise Semantic Platform is integrated to Microsoft’s FAST Enterprise Search Platform (ESP) providing:
- Model based navigation – drill up, down and across – to enhance the user search navigation experience.
- Improved precision and recall on search results by applying complex classification capabilities to add metadata to the FAST index.
Semaphore search components work alongside FAST’s facet navigation. Unlike FAST’s entity extraction that is based on algorithms and dictionaries of terms, Semaphore's navigation is based on a 'semantic model'. A semantic model can be a taxonomy, thesaurus, or ontology and could be developed in-house or might be used across an industry sector.
The Power of a 'Semantic Model' in Classifying and Organizing Information
Semantic models offer a way to organize knowledge and information. They make it possible to define a subject domain using a hierarchy of terms and the relationships between those terms. The model makes the subject clear to a user. These models help illustrate to a person what they know, what they don’t know and what they need to know. Whether researching a company; finding the right product; or identifying whom to talk to, the semantic model is a powerful tool by which people quickly find answers or stimulate problem solving.
Automatic Content Classification and 'Aboutness'
Semaphore applies the semantic model by automatic 'about-ness' classification of the content. The same information being loaded by the FAST indexing pipeline is passed to Semaphore Classification Server. Here complex 'rules-based' logic is applied to match terms from the taxonomy. This matching process only returns a tag if the system thinks there is enough evidence in the content to do so. A passing mention of the term will not be returned by Semaphore, but will typically be returned by other statistical and algorithmic classification mechanisms. Incorrect or inaccurate subject tagging is arguably worse than no taxonomy tags at all.