This website contains the corpus of all sentences of the German part of the USAGE corpus and the machine translated sentences (using the Google Translate API). In addition, quality measures of the translation have been added manually. A detailed description what this corpus has been used for is available in our CoNLL 2015 paper:
“Instance Selection Improves Cross-Lingual Model Training for
Fine-Grained Sentiment Analysis”
Roman Klinger and Philipp Cimiano, CoNLL 2015.
Please download the corpus here. To retrieve the password, send an email to firstname.lastname@example.org. We would appreciate if you provided a short note what you will use the data for. In addition, please state clearly in that mail that you agree to not redistribute the corpus or parts of it and only use it for research purposes.