Latent Semantic Analysis (LSA) and Search Engines (SEO)

By Jose Nuñez

Latent Semantic Analysis (LSA) is applied by taking millions of web pages, where the search engines can learn which words are related and which noun concepts relate to one another. Searh Engines are considering related terms and recognizing which terms that frequently occur together, maybe on the same page, or in close enough proximity. So it is mainly used for language modeling or most other applications.

Part of this process involves looking at the copy content of a page, or included on the links, and looking through the ways on how they are related. Latent Semantic Analysis (LSA) is based on the well known Singular Value Decomposition Theorem from Matrix Algebra but applied to text. That is why some of the semantic analysis that is done at the page content level it may also be done on the linkage data.

LSA represents the meaning of words as a vector, thus calculating word similarity. Iit has been very efficient to that purpose, and is still used. Regarding text for this application, is considered linear. This makes LSA slow due to using a matrix method called Singular Value Decomposition to create the concept space. But it does only address semantic similarity and not ranking, which is the SEO priority.

Scientific SEOs have a similar goal. They try to discover which words and phrases are most semantically linked together for a given keyword phrase, so when Search Engines crawl the web, they find that links to particular pages and content within them is semantically related to other information that is currently in their database. So, in conclusion, LSA calculates a measure of similarity for words based on possible occurrence patterns of words in documents and on how often words appear in the same context or together with the same set of common elements.

Latent Semantic Indexing and Search Engines Optimimization (SEO)

By Jose Nuñez

The closest search engines have come to actual applications of this technology so far is know as "Associative Indexing" and it is put in effect under Stemming, or the indexing of words on the basis of their uninflected roots (plurals, adverbs, and adjectival forms are reduced to simple noun and verb forms before indexing).

Latent Semantic Analysis (LSA) is a technique in natural language processing, in particular in vectorial semantics, invented in 1990 [1] by Scott Deerwester, Susan Dumais, George Furnas, Thomas Landauer, and Richard Harshman. In the context of its application to information retrieval, it is sometimes called Latent Semantic Indexing (LSI).

Here are some quick facts about Latent Semantic Indexing:
1. LSI is 30% more effective than popular word matching methods.
2. LSI uses a fully automatic statistical method (Singular Value Decomposition)
3. It is very effective in cross-languages retrievals.
5. LSI can retrieve relevant information that does not contain query words.
6. It finds more relevant information than other methods.

Latent Semantic Indexing adds an important step to the document indexing process. In addition to recording which keywords a document contains, the method examines document collections as a whole, to see which others do contain some of those same words. LSI considers documents that have many words in common to be semantically close, and ones that have few words in common to be semantically distant. This method correlates surprisingly well with how a human being looking at content, classifies multiple documents.