tedster - 4:49 pm on Sep 28, 2011 (gmt 0)
Tedster, can I ask what you mean by the above and how do you figure out which are these co-occuring terms? Is it based on competitor research?
"Co-occuring terms" is a long established concept in IR (information retrieval, the art/science that evolved into web search.) The idea is that there will be some words and phrases that commonly occur together in documents about any topic - think "doctor's office" and "blood test", for example.
Those two example phrases are "2-grams" - a phrase of "n" words long is called an "n-gram". Google's patents about phrase based indexing [webmasterworld.com] lean on these concepts intensively. They process their entire corpus of web page data regularly to identify the phrases that are in actual use, and to identify newly emerging language conventions.
The absence of co-occurring phrases in a document can be a sign of "over-optimization" - too much focus on just the targeted keyword phrase in the content. Likewise, the presence of too many can be a spam signal. Patching together a lot of scraped sentences from many websites often creates this situation.
Years ago, there used to be an online service (it's now gone) that took the top 50 results for any phrase and calculated the 2-grams, 3-grams and 4-grams that occurred in the top ranking documents. After using it a bit, I found that correlated phrases could be chosen rather intuitively by just poking around - reading the top pages, using various keyword research tools, the related search functions and so on.
For my purposes, I saw that I don't need the heavy duty calculations - just a handy jolt that pops my content writing out of the old-skool, lock step focus on "keyword phrase." In other words - more like the way normal, non-search-geek people would write about a topic.