Most respectable SEO companies understand the realities of LSI, but I still see company after company discussing LSI like they can control it and that it is a part of SEO. If you are unsure of what LSI is or if you offer it as a service I HIGHLY recommend that you read Dr. Edel Garcia‘s web sites!

• irthoughts.wordpress.com
• www.miislita.com

Here is an article I wrote last June referencing some of the top SEO experts in the country- Bill Slawski, Rand Fishkin and Lee Odden. Misconceptions & Misunderstandings of the LSI Algorithm

Here are some of Dr. Edel Garcia‘s tutorials:

SVD and LSI Tutorial 5: LSI Keyword Research and Co-Occurrence Theory
SVD and LSI Tutorial 4: Latent Semantic Indexing (LSI) How-to Calculations
SVD and LSI Tutorial 3: Computing the Full SVD of a Matrix
SVD and LSI Tutorial 2: Computing Singular Values SVD and LSI Tutorial 1: Understanding SVD and LSI
LSI Keyword Research – A Fast Track TutorialLatent Semantic Indexing (LSI) Fast Track Tutorial
Singular Value Decomposition (SVD) Fast Track Tutorial

 Melissa- SEOAware.com

Many people understand LSI as semantically close or distant and SEOs need to keep this in mind when optimizing. I believed this as well; I believed what I was told. Well guess what? This is totally wrong!

Dr. Edel Garcia has taken a lot of time to explain the misconceptions and misunderstandings of LSI. I suggest that anyone interested in SEO or those that have paid for LSI friendly services go to Dr. Garcia’s Site and learn about LSI.

Dr. Garcia and I discussed the issue and I asked him how I could explain LSI and the misconceptions easily or simply and he responded by email. His answer is below. I would like to apologize for posting an LSI blog here that was incorrect. I also hope that other SEOs that have paid money to educate themselves on SEO are as upset as I am that we were given false information.

———————————————————————————

”

I like to go to those new in the SEO industry and try to educate them against these purveyors of falsehood, spammers, or plain snakeoil marketers, they can go ahead and teach others. Those SEOs that are veterans at deceiving the public under the impression of being “experts” will not change their habits since that’s how they make their money.

Some of these unethical firms have been around since the days of Lycos, Infospace, Altavista, WebCrawler, and other early search engines from mid 90s. Within IR circles we just laugh out laud at their irrisory ideas on how actually search engines work. That’s a good reason for the W3C to have the special track on AIRWeb (Adversarial Information Retrieval on the Web): to debunk these spammers and their tricks.

See http://www.miislita.comfor the conference being organizing in Canada. I am a program committee member of this AIRWEB07 version. Regarding putting things in simple terms: Irronically, if you skip the math and look at the results and figures of the last tutorial of the series (#5)SVD and LSI Tutorial 5: LSI Keyword Research and Co-Occurrence Theory

it shows a simple example on how term weights in the LSI doc-matrix are redistributed across a small collection of just three docs. Now imagine a collection of 1million or 1 billion of docs. A large matrix, hugh? Thus, this explains why optimization strategies cannot affect LSI or the SVD for that matter, unless SEOs have access to the full index of the search engines and have superpowers to read the intention of other publishers or editors out there. To sum up: there is no such thing as “LSI SEO optimization”, “LSI-friendly”, or “LSI-compatible” documents. Same with LSI and link popularity. Pure non sense. In the example, I used a primitive model wherein term weights are defined as word occurrences in docs. This model was taken from Prof Grossman and Frieder’s book which I reviewed long ago.

The picture of computing term weights is actually more complex since no current search engines define term weights as mere occurrences (term frequencies), but using composite weights

aij = Lij*Gi*Nj where aij = term weights
Lij=local weights
Gi=global weights
Nj=normalization weights

Some even define Gi in terms of entropy weights, but I am not going to explain this now. I have tutorials explaining entropy weights and all these equations at Mi Islita.comThe thing is that each search engines have their own recipe for defining Lij,Gi,and Nj. Once aij is computed, these values are then used in the term-doc matrix to be decomposed in LSI.On and on, if you want to go ahead and blog, don’t forget to mention the following SEO articles. These authors learned from me the truth about LSI implementations and then went ahead and educated others on these topics:

Mike Grehan
Lies, Lies, and LSI by Mike Grehan
http://www.clickz.com/showPage.html?page=3623571

Bill Slawski
http://www.seobythesea.com/?p=535

Rand Fishkin
http://www.seomoz.org/blog/infosearch-media-contentlogic-purveyors-of-falsehoods

Lee Odden
http://www.toprankblog.com/2006/12/5-myths-about-seo/

Searchgrit
http://www.searchgrit.com/the-history-of-latent-semantic-indexing/