Statistical methods can be used to automatically generate a given domain's semantics system, by analyzing a representative document sample.
	Semantics can also be implemented by direct loading of existing dictionaries and thesauruses.
	A combination of the above two methods can also be implemented.