Topic-Attribute Relationships

This app attempts extract basic information from text inputs. Topics and their attributes are extracted. The number of times an attribute is found for a given topic is counted, once per paragraph.

A suitable use case might be a list (one response per return-ended "paragraph") of open responses on a questionnaire.

Input

Text to analyze:

Submit

Example

An example to check whether the expected topic-attribute relationships are returned:

Text to analyze:

I like dogs, but I don't like cats.
Dogs are great. Thomas eats fish and is smelly. Jim is suspicious.
Cats are evil. Thomas is smelly and quick.
I like hats. Hats are the best. Cats are definitely evil.
I feel cats are evil and Thomas is smelly.
The evil of cats astounds me, but dogs are great.

This should return lists of topics, attributes and their associations that reflect the gist of the statements.