Semantics

Latent Semantic Indexing - What Is It And How Can You Benefit?
By Paul Hooper - Kelly

What is Latent Semantic Indexing? It certainly sounds impressive and mysterious. Yet, like most jargon, it's quite logical, once you understand the meaning. More to the point: once you understand it, you can use it to boost your search engine listings.

"Latent" means "hidden" and "semantic" stands for "meaning". So LSI is simply "hidden meaning indexing" and it's a phrase you will come to hear more and more now, in relation to search engines -- particularly Google.

Google, having set the standard for search engines -- and shown the others how they have monetarized their store of information, with their AdWords program -- now find both Msn and Yahoo snapping at their heels.

But Google didn't get to be top dog in the search engine wars without keeping on the cutting edge of information retrieval technology, which is why they have been quietly acquiring technology and knowledge for the last few years, even to the extent of buying complete companies. This technology is already being used and will be increasingly used to ensure Google stay ahead of the competition.

What brought Google to such prominence from a standing start, seven years ago, to the lion's share of the search engine market was their strict adherence to the twin mantras of "relevance" and "quality".

If you use a computer, you'll understand the paradox: they provide of being amazingly clever for some functions and totally stupid for others. So, whilst Google's computers were pretty good at checking relevancy -- for example matching keywords to the text of the web pages -- they have been pretty poor at measuring the quality of the web page.

So, to gauge this, they had to rely on information they could get a handle on, such as links to a web page from sites run by humans. The logic cannot be faulted: if a human, thought the content was good and signaled their approval by sending a link to a web page, Google took this as a vote for the content. Several links from several different web sites was even better.