Domain-Specific Term Extraction: A Case Study on Greek Maritime Legal Texts.

Preservation of cultural heritage has significantly attracted many research efforts. Amongst them, significant interest has been presented to sharing of cultural heritage content, whereas one major aspect of cultural content is the one concerning maritime heritage. In order to efficiently consume such content, high-quality descriptions are required. Towards this goal, we propose an approach for automatic term extraction of documents, where our methodology is based on the use of word embeddings along with a deep neural network architecture. We demonstrate its efficiency by using a large corpus of legal documents related with maritime.