Latent Semantic Indexing

First some basic understanding of language ….

Synonym

Defined as “word or phrase that means exactly or nearly the same as another in the same language”.

For example one synonym of “argue” would be “reason”. If I do a search on Google for “argue”, the results are all related to “argue” - argument, how to argue, etc.

Polysemy

- the ambiguity of an individual word or phrase that can be used (in different contexts) to express two or more different meanings

For example: Mole as in 1) Burrowing animal and in 2) Espionage.

How does this apply to Google?

If you contemplate a search on Google for “apple”. What would you expect to see? Items related to Apple with the meaning of fruit?

Well - do the search and what is the result? The first page of results all relate to Apple computer and its products and services!

But Google is intelligent to know that if you expand the phrase to “apple tart” - then the search results relate to fruit pie rather than an Apple Computer slut!

What is the significance of these observation?

Well if I am creating a web page or site on Apple (as in computers), I would not be using a Thesaurus to find alternative words for “Apple”. Why? The Thesaurus will produce variations on Apple the fruit! .

So what is important is understanding Google’s Thesaurus which IS based on LSI. Google will highly rank sites who are authorities about that Keyword or phrase. How does Google use LSI to work that out? With Google’s resources (online libraries of books, etc) it has worked out what associated terms and phrases you would expect to find on the page (or website) to reinforce the search term.

Example of LSI at work

For example at its simplest level, for “Apple” through its LSI technology it expects to find the terms Apple, Apple Computer, Quick Time, Windows and Mac on the page! How do I know that? See Tools

Next >> How does LSI apply to your site

Popularity: 100% [?]