Android Question NLP - Questions based on the Paragraph... B4A

[StephenRM]

Consider the passage below:

" Thanappa, a postman who always talks to the people on his route. He delivered to a man named Ramanujam from before his daughter was born to the time she is of marrying age. After the last marriage prospect falls through, Thanappa recommends a suitor in a faraway town. The meeting goes well and everyone is rushing to get the wedding done May 20th. If it is delayed even one day, the groom's "training" will delay it three years. After the wedding, Thanappa reveals a postcard to Ramanujam about the death of a relative, which should have been delivered May 20th. Thanappa did not want to disrupt the wedding, even though hiding the letter should cost him his job. Ramanujam says he will not issue a complaint. "​

(smal extract of a story from RK Narayan's, Malgudi Days)​

Is there any way to frame questions on the above paragraph as;
1. Who was Thanappa?
2. When was the wedding done?
3. What did Thanappa reveal to Ramanujam?

Thanks and Regards
Stephen

 

[Brian Dean]

My first thought was "No", but then I thought again. As a simple example, one could scan for the word "is" and the word structure "something IS something". Taking this passage as input this rule could lead to two questions ...

- Who is of marrying age?
- What is delayed even one day?

Obviously the second question is not properly formed, but transposing the word "she" to "Who" and the word "it" to "What" are not too complicated. So to answer the question "Is there any way to frame questions on the above paragraph" I would say "Maybe - but it is not going to be easy."

[Edit - I missed a third question - "Who is rushing to get the wedding done?"]

[Second edit : Since my first reply I have, with the help of Google, looked for a more constructive answer. I get the impression that this area (semantic analysis and automated comprehension testing) is more an academic than practical topic at the moment. I found plenty of scholarly papers but no APIs (unless you want to analyse sentiment from customer feedback.)]