Indexed algorithms for approximate structure matching



There are many papers on Google Scholar on indexed algorithms for
approximate STRING matching, indexed algorithms for approximate TREE
matching, etc.

I wonder whether we could design a general framework for treating
general cases of such problems, that is indexed algorithms for
approximate STRUCTURE matching.

This would make it possible to tackle a wide range of application
problems without dedicating research resources to each and every of
them:

* Finding literally nearly identical web pages so that we can combine
their reader comments. A nice feature suggestion for the recent Firefox
extension Blogger Web Comments launched by Google
(http://www.bytecool.com/temp/google-webcomments.xpi).

* Source code / essay plagiarism checkers;

* Shape search; content-based image search; voice recognition;
character recognition;

* Matching online people and knowledge literature of most similar
special interests; (expressions of thought can also be viewed as
structures of semantic concepts)

* Scientific data search (e.g. DNA sequences)

Regards,
Yao

.