Monday, October 6, 2008

My Research Paper

Hi everyone,
I would like to share my idea of a better search engine with all of you, I am posting the abstract, you can read the whole paper here
ABSTRACT

The Next Search Engine

ABSTRACT
Ubiquity of internet has opened doors to affluence of knowledge. However, the enigma to find your needle in the haystack still fang out invincibly. Most search engines implements keyword based search which although popular is inefficient as it excludes the desired web pages on the basis of absence of keyword(s).User is prone to browse useless pages as keyword based search generates many false positives and fails to produce the set of web pages that have significance in terms of the context being searched.

We have proposed an approach, the Next Approach, which handles variation in perception of the human user by enhancing the pool of ‘relevant’ web pages. Being a Neighborhood-based strategy, it searches for a word or links that lie in the neighborhood of the keyword in terms of some contextual similarity measure with the given keyword. The approach involves identification of some other word that is similar in some context with the present keyword. To do this, the content of URLs containing the keyword are scanned to identify the next most relevant keyword. Working on the similar track, a scan is performed to identify the referred URLs and made a judgment about a set of URLs that might be relevant in terms of the context, even though they did not contain the keyword.

1 comment:

Kazim Zaidi said...

Did you try the cuil search engine?

It is content based, plus a lot of linguistics that goes inside to find the pages with right context. They analyze not just keywords, but the coherency of the entire page (keywords viewed as a whole).

That means, content-based searching with linguistic analysis. And not based on link structure (so no more affiliate hijacking that people do for Google).

And they claim to have thrice as big a database than Google. :-P Check it out!