如何从文章中提取（识别）书名？

Question

Is there any good method to extract (recognize) book title from the article using nltk or something else?

I can recognize author names using nltk, so my idea is to get list of book titles with authors from some external source and when I recognize author name then I could take a list of this author's books from external source and look for them in the text.

but I'm not convinced about this solution because I need external source with all books and I don't have such a source and this solution seems a bit like "brute force" for me.

can you direct me on topics that will help me with this problem?

Answer 1

Given sufficient training data, there is a wonderful python library for achieving things like this called https://github.com/snipsco/snips-nlu

What you might want to do is grab examples from as many articles that include book titles as you can, follow the documentation on that repository, and you should be able to glean book titles from articles assuming they follow a similar pattern as your example data.

I'm not 100% positive that this is a task for machine learning however. There may be an easier way, such as looking for words/phrases that are in quotes, are italicized, etc. Humans don't necessarily know that a bunch of words are the title of a book, so we invented punctuation to explicitly make that clear. It seems to me there should be some use for that syntax in your solution if possible.