optimaize / language-detector

Language Detection Library for Java
Apache License 2.0
567 stars 165 forks source link

Upper case English text returns a low probability #90

Open vhothur opened 6 years ago

vhothur commented 6 years ago

Performing a getProbabilities with text in upper case, returns a low probability score for English (0.4-0.5 range) and almost always returns a probability score for German but lower (around 0.1-0.2). Changing to lowercase does returns the correct probability (0.99)