The Discovery Engine approaches search in a completely different way. The Discovery Engine uses its ability to understand relationships to calculate the relevance of all objects to one another. When a user submits a query, the engine examines the relevance of all oabjects in the data set to the query and reorders the data set from most to least relevant.
“Fuzzy” search is the process of showing the most relevant results in a data set-the best matches- regardless of whether the results exactly match the criteria a user is searching for.
Advantages of Fuzzy Search over Traditional Database Search
Fuzzy search allows users to be specific about what they are looking for. The Discovery Engine will return the best matches for even the most detailed queries. We believe that users should be able to engage search systems intuitively like a human being, not literally like a computer.
Fuzzy matching requires large numbers of complex calculations. Traditional database systems cannot handle the performance implications of fuzzy matching at scale. We architected the Discovery Engine from the ground up to power fuzzy matching over millions of structured data and free text records and hundreds of metadata fields at best-in-class speeds.
No guessing required.
Traditional search systems are not fault tolerant. They require users to guess at contents of the database, send a query to the database, see if any satisfactory results come back, and, if not, repeat until they do get satisfactory results or give up trying. The Discovery Engine removes the guesswork. Users enter a single query and all the results are returned in order of relevance.
+ 1 (212) 982-9887 email@example.com
Eliminates “No Results Found” error.
The irony inherent in traditional database search is the more information users give about what they’re looking for, the smaller the result set that gets returned. Fuzzy search always returns the maximum number of results, and the quality of the results increases as users are more specific in their requests.
Traditional search systems are unidirectional: the criteria sought by the search initiating party are compared with the criteria offered by other parties. Often, however, bi-directional matching-in which both parties’ desires are considered-is more preferable. For instance, job candidates have preferences for certain kinds of positions, and companies filling those positions have preferences for candidates with particular qualifications. Unidirectional search fails to take into account the reciprocal preferences of both candidates and employers, resulting in higher search transaction costs.
Show similar results.
The Discovery Engine uses fuzzy matching over multiple metadata points to calculate how similar records are from one another. This functionality has many potential uses. For example, an e-commerce site could use fuzzy matching to provide users with a “Similar Products” widget. An online real estate site could employ fuzzy matching to improve the email notifications of new listings that they send to users.