Try this search. “Find digital camera stores near me”. Guess what I found when I typed this into Google? I got a bunch of links on “near infrared cameras”. Yahoo and Microsoft search did no better. That is because these search engines, look at keywords in your search and try to find a match. They try to match your search terms to keywords found on web pages.
There are two problems here. One is that search engines do not get the intention of the search. The second problem is that the web pages do not provide any clues to the search engines for locating pertinent information. Search engines not only have to guess the intention behind your search, but also infer information from the mass of text on web pages.
If I am a smart digital camera shop owner, I may want to embed some special information in my web pages to show that I sell digital cameras. In addition, I may choose to provide some location information too. There are several techniques for embedding more meaningful (semantics) information on web pages.
An informal tagging technique called Microformats is one of those. A recent effort called Embedded RDF by W3C is another. These approaches allow you to tag information on your web pages. If you are selling any kind of services or products, you should seriously look at semantic tagging using Microformats or Embedded RDF. You may want to consider this for the next revision of your website. You can start small, tag a few pieces of essential information.
2 thoughts on “How Can We Make Search Better”
Even if the pages contain additional information
the conundrum is how the search engine understands
the meaning of the user supplied query.
One such effort is there @ powerlabs { http://www.powerset.com/ }
(powerset) search engine;
it’s not yet been released for public use.
but they are providing it for limited set of people for evaluation;
me too had spent some time with that;
seems to be interesting….
Thanks Kuppu. I signed up for Powerset. I will check it out.
There are lots of smart things search engines can do. Key word search is just one aspect of search. There is something called “Concept Extraction” in which a program can read and extract the concepts and relationships in a given body of text (based on a set of pre-defined rules and AI). My knowledge in this area is mostly peripheral.
You may also want to check out the AIML chat bot called Alice (http://www.alicebot.org/). It is amazing how a simple rule driven (in this case artificial intelligence markup language) based program can mimic almost human like behavior.
Comments are closed.