Lately I’ve been asked what the implicit phrase matching exactly means. The answer is pretty long so cannot fit into a tweet so I decided to put it into a blog post…
As you know, search functions are based on two major components: crawling/indexing and querying. Query server receives the user queries, build the answer based on the indices, and send it back to the user. The result set contains a set of results, ordered by the ranking of each of them (in FAST Search Server for SharePoint 2010 you can change the sorting property). Ranking algorithms can be very complex, but basically they have some common things. For example, the most important factors of a term’ rank are the frequency and location of keywords.
Frequency means: how many times the keyword can be found on the page. The location of the word is also important: for example, keywords in the Title or Head tags of the page are more relevant than the words in the text only. But location can also mean the neighborhood of the keyword: what other keywords are near to this.
This is where phrase matching shows up: explicit phrase matching means that an explicit user-defined condition will be evaluated, for example the query SharePoint NEAR administrator is looking for documents where these keywords are close to each other. Implicit phrase matching is the process, when the keywords’ ranking next to each other will be increased and the results show up at the top of the result set (for example SharePoint administrator).
Leave a Reply