eTRAP’s article “Sentence Shortening via Morpho-Syntactic Annotated Data in Historical Language Learning” authored by Maria Moritz, Barbara Pavlek, Greta Franzini and Gregory Crane, is now published in the current issue of the ACM Journal on Computing and Cultural Heritage (JOCCH). The work was supported by the Federal Ministry of Education (BMBF) and the European Social Fund (ESF). Here is the abstract:
We present an approach to shorten Ancient Greek sentences by using morpho-syntactic information attached to each word in a sentence. This work underpins the content of our eLearning application, AncientGeek, whose unique teaching technique draws from primary Greek sources. By applying a technique that skips the clausal dependents of a main verb, we reached a well-formed rate of 89% of the sentences.