Algorithm to suggest a list of tags to users

Question

Given a free text, I need to analyse this this text and suggest a list of tags from a pre existing list.
What algorithms are out there in the market? Can they handle a case where, for example, the text have a word like

high cholesterol

and I would like it so suggest

heart disease

although "high cholesterol" might not exists (initially) in the pre defined list.

Answer 1

Sounds like you need a many-to-many relationship between key-words (free text) and your tag list. This way you can enforce all of the correct associations.

The computer itself will not be smart enough to know that high cholesterol and heart disease are related without someone defining the relationship in a database first.

Answer 2

Have a look at this article:

http://en.wikipedia.org/wiki/Association_rule_learning

Maybe it helps