Would it be possible to use wolfram language to segment unstructured text into a defined set of classifiers. For example, given an address the ability to identify what parts of address refer to house number, roads, city, state etc? If yes, what documentation should I be referring to?
Take a look at GrammarRules, Classify, and Strings and Characters. And a million of other things - just search in docs.