Forum Moderators: open

Message Too Old, No Replies

Keyphrase Extractor

how can we do it?

         

xzqi

12:08 pm on Dec 29, 2001 (gmt 0)



I am a newer here and I appreciate the generous help offered.
when we search the phrases in the search engines, the result will offer the pages including the phrases.how can the engines work? do they use a phrase dictionary to index the pages? or do they merely offer the pages in which the single terms of the phrase appear? I mean what do they index the pages using? the phrases dictionary or terms dictionary?
if the first method is used, do someone know about such an dictioanry to extract the keyphrase?

Air

6:28 pm on Jan 1, 2002 (gmt 0)

WebmasterWorld Senior Member 10+ Year Member



Welcome to WebmasterWorldm xzqi!

>how can the engines work?
The engines' method of displaying search results is based on that engine's algorithm of how they analyze a page to determine for which keyword searches it will be relevant and therefore returned in the results page shown.

>do they use a phrase dictionary to index the pages?
They use the content on your pages, some of the phrases come directly fom the pages they have indexed. In more advanced algorithms routines make associations between the page content, the overall content of multiple pages at the same site, links from sites that link to you, alternate meanings for keyphrases on your pages, and even context meanings, the list goes on and on, suffice it to say,the exact way the engines do this is not disclosed, it is a closely guarded secret by each engine.

>do someone know about such an dictioanry to extract the keyphrase?
You might be thinking about this in reverse. There isn't a finite set of keyword phrases to which the engines match sites, it is the reverse. A finite set of sites is returned based on a search phrase invented by the surfer.

Brett_Tabke

2:32 pm on Jan 3, 2002 (gmt 0)

WebmasterWorld Administrator 10+ Year Member Top Contributors Of The Month



Also, they see thousands upon millions of searches per day. In a very short time, they can build up lists of top searched for phrases. When they index and rank a page going into the database, they can create key files for that phrase that link back to that listing in the index. In that sense, a search engine works in reverse by indexing the phrase first and then finding a page that matches it.

xzqi

2:20 pm on Jan 4, 2002 (gmt 0)



At first thank you two for your warm response.I do acquire a lot!
but I am still confused by something.Since the pages are not indexed in the database,when users offer their search, how can the engines work?Does it first use the keyphrases given by users to index(maybe here "match" more suitable) the pages? It's so unefficiet!Then without a dictionary,how can it determine the relevance between the keyphrase and the pages?