Unmatched search services
Combining both statistical and semantic analysis, MatchPoint 3.0 from TripleHop Technologies is said to be new enterprise contextual search software designed to give corporate users fast and precise access to knowledge assets relevant to their job functions. Further, says TripleHop, MatchPoint 3.0 gives users increased sensitivity to the context of their query, including the user profile and existing enterprise knowledge.
The company explains that MatchPoint automatically generates a contextual thesaurus as it crawls and indexes enterprise documents. It uses this thesaurus to "disambiguate" and expand the user query based on existing enterprise knowledge. MatchPoint also takes into account the user profile and the customer's work process, acknowledging the context of the task in order to understand, disambiguate and expand a query to obtain more relevant results faster, says Triplehop.
MatchPoint 3.0 uses a distributed architecture to create as many "search services" as necessary. Each service can have several indexes, and each index can be partitioned. User queries are launched into multiple indexes and sub-indexes, and the results are merged through the MatchPoint Federator. This distributed architecture helps handle terabytes of data without substantially degrading response time. For example, response times range from 85 ms for an index size of 4.5 GB to 950 ms (0.9 second) for an index size of 70 GB, the company reports. TripleHop adds that a document can be indexed and ready for search within 10 seconds after it has been received or created.