Latent Semantic Analysis is a mathematical technique for analyzing data and patented in 1988 by Scott Deerwester. The term Latent Semantic Indexing (LSI) is a system to rank documents depending on the content. LSI is successfully used by search engines to detect spammers and poor content developers who try to confuse the engines by introducing a volley of keywords just to obtain higher rankings.
LSI provides a freedom from using repetitive words. Using more than one synonym to make the content look varied and without monotonous sounding words has been enabled. The use of polysemous words has also been made easier. LSI takes into account the subject of the document and decides the appropriate meaning of the word. Using keywords has been simplified by LSI. When searching for keywords on Google's free tool, the related words that also can be used as keywords are shown. This can be very helpful in designating a widespread of keywords. LSI has promoted the variability in vocabulary used in contents with respect to phrases which can be replaced by a single word and vice versa. For example, the phrase "penal labor camp" or "concentration camps" mean "Gulags" or "Nazi camps"
Exploit the creativity that is allowed by LSI. LSI helps the search engine in identifying the exact meaning of a given word when used in context of the sentence. This does away with the need to introduce the particular key word randomly throughout the content. LSI also helps the engine relate keywords with phrases that may not contain the word but are very much relevant. An example can be the results thrown up if you put Saddam Hussein in the search box. The 1982 massacre in Dujail and the 1988 massacre in Hallabja are shown at the highest rankings. But, also shown are relevant topics related to Iraq and the UN sanctions in which his name is not mentioned. This can be exploited to relate your content to other sites and pages which may have the same context even though the keywords are different.
LSI immediately helps in finding relationships. Hence, while writing you need not add the exact words or their synonyms every time. This opens up space for creativity without actually deviating from the subject.
Do not build content which revolves around only one keyword. LSI is used by search engines to value content when viewed as a whole. Such monotonous content may be mistaken for spam and rankings may be lowered.
While developing content with words with many meanings the context should be clearly mentioned. A couple of examples will drive home the point. While using the word "Kiwis", a clear mention should be made about whom are you referring to - the bird or the people of New Zealand. Another example can be the "tension" - meaning low tension wires, psychological distress, cables for bridges, etc.
LSI is very useful in identifying duplicate articles circulated on the web. Hence, be very careful in preparing content with a unique quality. Duplicity can also be overcome by changing the view point of presentation and relating the content to current events and ideas. Consider an article related to dating. If the content is just copied from the web, and after alterations in structure is displayed under a new title, LSI can detect the duplicity. A better way would be to add research analysis of the latest happenings and events in the world of dating. This could include a new dating site or blog that has appeared on the web. This vital string of up to date information can improve ratings, in spite of dating being a common subject.