HORUS is a Named Entity Recognition Algorithm specifically designed for short-text, i.e., microblogs and other noisy datasets existing on the web, e.g.: social media, some web-sites, blogs and etc..
It is a simplistic approach based on multi-level machine learning combined with computer vision techniques that aims at provide metadata for each token of an input sequence. Example:
#!/usr/bin/env python from horus.components.core import Core
horus = Core() x = horus.annotate("j bond was born n glencoe") print x
more info at: https://github.com/dnes85/components-models
# Author: Esteves <diegoesteves@gmail.com> # Version: 1.5 # Version Label: HORUS_NER_2016_1.5 # License: Apache