RDFFrames: knowledge graph access for machine learning tools

Abstract Knowledge graphs represented as RDF datasets are integral to many machine learning applications. is supported by a rich ecosystem of data management systems and tools, most notably database that provide SPARQL query interface. Surprisingly, tools for knowledge do not use SPARQL, despite the obvious advantages using system. This due mismatch between in terms model programming style. Machine work on tabular format process it an imperative style, while declarative has its basic operation matching graph patterns triples. We posit good interface from software stack should imperative, navigational paradigm based traversal rather than patterns. In this paper, we present RDFFrames, framework provides such RDFFrames Python API gets internally translated integrated with PyData stack. enables user make sequence calls define be extracted stored system, translates these into compact SPQARL query, executes returns results standard format. Thus, useful tool preparation combines usability flexibility performance systems.