A filtering proposal for extracted Arabic term candidates

In the terminology extraction process, determining relevance of the candidates is very crucial for the purpose of identifying domain-relevant terms. Information about terms can be often gathered from linguistic knowledge or from statistic measures. In this paper, we present a proposition of a filtering mechanism based on a machine learning technique so as to keep only the most relevant terms. The proposed strategy incorporates varied and rich features from the content as well as the structure of Arabic legal documents.