Provenance in Data Science

<div>RDF-based knowledge graphs require additional formalisms to be fully context-aware which is presented in this book. This book also provides a collection of provenance techniques and state-of-the-art metadata-enhanced provenance-aware knowledge graph-based representations across multiple application domains in order to demonstrate how to combine graph-based data models and provenance representations.  This is important to make statements authoritative verifiable and reproducible such as in biomedical pharmaceutical and cybersecurity applications where the data source and generator can be just as important as the data itself.</div><div>             </div><div>Capturing provenance is critical to ensure sound experimental results and rigorously designed research studies for patient and drug safety pathology reports and medical evidence generation. Similarly provenance is needed for cyberthreat intelligence dashboards and attack mapsthat aggregate and/or fuse heterogeneous data from disparate data sources to differentiate between unimportant online events and dangerous cyberattacks which is demonstrated in this book. Without provenance data reliability and trustworthiness might be limited causing data reuse trust reproducibility and accountability issues.</div><div><br></div><div>This book primarily targets researchers who utilize knowledge graphs in their methods and approaches (this includes researchers from a variety of domains such as cybersecurity eHealth data science Semantic Web etc.). This book collects core facts for the state of the art in provenance approaches and techniques complemented by a critical review of existing approaches. New research directions are also provided that combine data science and knowledge graphs for an increasingly important research topic.</div>