The term “Latent Semantic Indexing and Analysis” (LSI/A) translates into “hidden meaning”. Search engines, primarily Google, are using natural language processing to determine relationships of set terminology. Using a term called Singular Value Decomposition, search engines can relate terminology to the theme of your site. LSI/A can find relationships between terms used on your site. Without a plethora of terms all related to an overall theme of your Website, a search engine would fall into two fundamental problems called Synonymy and Polysemy. Synonymy is where two different words express the same idea (i.e. intelligent | smart). Polysemy is where the same word has multiple meanings (i.e. apple | Apple ). A search engine will use the related terminology of your site to pinpoint the exact meaning.
Phrase Based Indexing and Retrieval (PaIR) is a similar technology used for Spam detection. It calculates a Spam threshold by calculating topical phrase occurrences. If a keyphrases and keywords are related in an overall theme then your site won’t trigger a Spam filter.
Latent Semantic Indexing (LSI), Latent Semantic Analysis (LSA), Phrase Based Indexing and Retrieval (PaIR), vector analysis, and singular value decomposition are all very technical issues involved in algorithmic search. Save yourself from rigorously analyzing Google’s patents and studying the mathematical equations involved , and let me dumb it down as simple as possible: WRITE YOUR CONTENT FOR YOUR USERS, AND HAVE A CENTRALIZED THEME. As long as you keep your content relevant and you’re not trying to game the search engines you won’t have to worry about LSI/A or PaIR filtering your Website.