Given a longer English text (> a few paragraphs), is there a rule-based NLP approach to classifying a set of text to be occurring at a place or time, from an establishing phrase? For example:
Alice went to London. She met Bob at his hotel, and they went out to dinner. They discussed old times and made a new friend, Cassandra. Over the next week, ...
[more paragraphs of activity, implicitly in London]
Alice then left London and went back to Madrid. She...
As readers, we can say that Alice, Bob, and Cassandra all were in London and performed various actions there. But I can't find how to do automate this, how to match NLP techniques that I have read about with this more specific task.
Are there current tools or research that deal with this problem? I am interested in unstructured, likely narrative, text. Even a name for what this is would be helpful, something more specific than "text mining" or "text classification".
I understand that there will not be 100% accuracy with a rule-based approach. I also understand a sufficiently trained algorithm would perform this task better, but I am interested in something quick and dirty right now. Are there out-of-the-box tools that do similar work, rule-based or trained? (I am enjoying StanfordNLP, and the OpenIE module is pretty good, but doesn't quite get to this specificity, afaik.)
