I'm creating a model/system designed to extract topics said in a corpus/text. But first I realized it's best to first make a pass on the corpus to know what it's talking about (its domain ).. Then extract the topics in relevancy with the domain. For example: text talking about math/geometry -> detect domain is math-geometry -> extracted topics triangles and acute angles, ... (what the text contained) Is this doable? If so, how? I don't know if...
Originally asked by helge1953 on Stack Exchange (ai) (CC BY-SA 4.0).
