Scientific Highlight “Open Information Extraction for the German language”

Knowledge Discovery

GerIE is the first grammatically based German Open Information System which identifies factual statements in written text purely based on the grammatical structure.

Based on feedback from company partners we studied the applicability of OIE methods to German. The task is to identify factual statements in written text purely based on the grammatical structure. A main outcome of this work is the first grammatically based German Open Information System not based on rules or models (GerIE). Also, Machine Translation techniques supporting OIE have been developed. These achievements have many applications: sentiment detection (Rexha, 2016b)KC, aspect classification (Falk et al., 2016)KC, patent analysis (Pimas et al., 2016b)KC, author writing style detection (Rexha et al., 2016aKC, Rexha et al., 2017KC), and scientific publication mining (Klampfl et al., 2016cKC, Felber & Kern., 2017KC). For the latter, work on PDF extraction was of help (Klampfl & Kern, 2016aKC, Klampfl & Kern, 2016b KC).