Deep Learning and Medical Data Engineering for Knowledge Discovery

Today we are facing an unprecedented number of created data, a vast stream of huge data but of questionable relevance. Data engineering has focused on big data - storing and processing large amounts of data. There is a potential big value in creating knowledge from this data. Deep learning is intended to create algorithmic expertise in a determined field, to narrow the gap between data and knowledge. This line of research has matured a lot in recent years, with the creation of tools and techniques to help data scientists achieve this goal. In this paper, we develop and confirm the hypothesis that data engineering is modulated by the application domain, called Domain Driven Data Engineering. The behaviors and choices of a data engineering process are mainly related to the domain itself, leading to a new characterization of data engineering activities and related concepts. As a result, we conclude with a new roadmap of data engineering phases accordingly to the Domain Driven-based methodology, refining previous related works in this field.