Dependence language model for information retrieval by Jianfeng Gao , Jian-yun Nie , Guangyuan Wu , and Guihong Cao, is a good introduction to dependency analysis in information retrieval.
The theory is that terms (words) in a document depend upon other words and that those dependencies can be used to improve the results of information retrieval efforts.
Beyond its own merits, I find the analogy of dependency analysis to subject identification interesting. That any subject identification depends upon other subjects being identified, whether those identifications are explicit or not.
If not explicit, we have the traditional IR problem of trying to determine what subjects were meant. We can see the patterns of usage but the reasons for the patterns lie just beyond our reach.
Dependency analysis does not seek an explicit identification but identifies patterns that appear to be associated with a particular term. That improves out “guesses” to a degree.
Topic maps enable us to make explicit what subjects the identification of a particular subject depends upon. Or rather to make explicit our identifications of subjects upon which an identification depends.
Whether the same subject is being identified, even by use of the same dependent identifications, is a question best answered by a user.