A simple topic detection module, based on Wordnet Domains data
A basic constructor
Returns a list of topics with normalized scores. There isn't a reliable order as JavaScript doesn't allow that, so you'll need to sort by score yourself to get the top scored topic.
var TopicDetection = require('topic-detection');
var detector = new TopicDetection();
var scores = detector.topics(myText);
Available under the MIT License. This module also contains data based on the Wordnet Domains data, used under the Creative Commons Attribution 3.0 Unported License.
Publications or presentations containing research results obtained through the use of WordNet Domains should cite at least one of the following references:
-
Luisa Bentivogli, Pamela Forner, Bernardo Magnini and Emanuele Pianta. "Revising WordNet Domains Hierarchy: Semantics, Coverage, and Balancing", in COLING 2004 Workshop on "Multilingual Linguistic Resources", Geneva, Switzerland, August 28, 2004, pp. 101-108.
-
Bernardo Magnini and Gabriela Cavaglià. "Integrating Subject Field Codes into WordNet", in Gavrilidou M., Crayannis G., Markantonatu S., Piperidis S. and Stainhaouer G. (Eds.) Proceedings of LREC-2000, Second International Conference on Language Resources and Evaluation, Athens, Greece, 31 May - 2 June, 2000, pp. 1413-1418.