I Would Like to know that Which algorithm is used for Keyword search?
The basic plan is victimisation AN inverted index. this implies for every word keeping an inventory of documents on the net that contain it.
Responding to question|a question |a question} corresponds to retreival of the matching documents (This is essentially done by decussate the lists for the corresponding query words), process the documents (extracting quality signals such as the doc, question pair), ranking the documents (using document quality signals like Page Rank and question signals and query/doc signals) then returning the highest ten documents.
Here square measure some tricks for doing the retrieval half efficiently:
- distribute the full issue over thousands and thousands of machines
- hump in memory
- wanting 1st at the question word with the shortest document list
- keeping the documents within the list in reverse pagerank order so we are able to stop early once we discover enough smart quality matches
- keep lists for pairs of words that occur oftentimes along
- sherd by document id, this manner the load is somewhat equally distributed and therefore the intersection is finished in parallel
- compress messages that square measure sent across the network
Jeff Dean during this nice speak explains quite few bits of the inner Google infrastructure. He mentions some of the previous ideas within the speak.
He goes through the evolution of the Google Search Serving style and thru MapReduce whereas giving general recommendation concerning building giant scale systems.
A search engine algorithm uses keywords as the input problem, and returns relevant search results as the solution, matching these keywords to the results.
|Find Web Hosting|
|Shared Web Hosting||UNIX & Linux Web Hosting||Windows Web Hosting||Adult Web Hosting|
|ASP ASP.NET Web Hosting||Reseller Web Hosting||VPS Web Hosting||Managed Web Hosting|
|Cloud Web Hosting||Dedicated Server||E-commerce Web Hosting||Cheap Web Hosting|