/topics

Predict news article topics and DBpedia description topics and type.

Primary LanguagePythonApache License 2.0Apache-2.0

This is a fork of https://github.com/KBNLresearch/topics/,

with an added category, music.

Model test output

Topics

Number of multi-class classifiers for news articles (topic) and DBpedia descriptions (topic and type). For further info and training source code see Jupyter notebooks for topic and type.

Web interface

A Bottle web application can be started with ./topics.py accepting requests with a newspaper article or DBpedia url as parameter, e.g. http://kbresearch.nl/topics_music/?url=http://nl.dbpedia.org/resource/Albert_Einstein. The application will work out the type of resource and language for itself and return all available classifier results.