To cluster linguistic patterns

[Moe]

Hello everybody
In my work, I would like to extract infomation from natural language text.
I have to find a set of extraction rule for my domain using a set of training text.
After finding extraction rule, I would like to cluster the extraction rule that have the same linguistic pattern.

For example, consider the two sentences
"<earthquake> has struck off < Java Island >"
and
"<earthquake> hit at <8:52>"


The two rules that represents the above sentences are grouped in the same cluster because they
all have the same linguistic pattern of
"<subject> active-verb preposition < direct-object>"

By using the parser I can identify their linguistic pattern.
To cluster the rule that have the same linguistic pattern, is there any clustering methods to be more faster that process?
I mean such as k-means algorithm.
Sorry if it is a stupid question.


Any help is appreciated.