machine learning - Topic Modeling Corpus

Question

As Im diving into topic modeling, I dont understand the process concerning learning corpus and the individual corpus of the documents you want to work with.

• Are they the same corpus? No right?

• So do you train the model on Wikipedia e.g., get topics and document distributions of wikipedia & later apply it on your book collection about whatever to find their topics?

• As you preprocess the data, do you preprocess both corpus if they are not the same?

• Do you evaluate the wikipedia learned model or the model/ topics of your individual corpus of books?

Thank you in advance!

Answer 1

等待大神解答