VoxEL.zip

This dataset has manual annotations respect to Wikipedia over the same text written in the languages: German (de), English (en), Spanish (es), French (fr) and Italian (it). This This dataset describes the information of composed of 15 news that has the same number of sentences in each language respectively, as well as the same number of annotations for each same sentence in the different languages. Each language has in general 94 sentences. We propose two annotations version of the datasets, one relax that include the maximum number of annotations and the strict one only those that are persons, organizations and places. The relax version have 674 annotations and the strict one only 204 annotations.