Graph Database Research, University of Bern (open questions)
Manuel Grutsch
Hi Guys
We are two students from the University of Bern, Switzerland. We are currently doing a comparison study of graph database management systems concerning the storage of fuzzy cognitive maps (cf. https://en.wikipedia.org/wiki/Fuzzy_cognitive_map). Fuzzy cognitive maps are a widely recognized technique to model knowledge entities and the relationships between them. However, there is currently no possibility of a large scale storage of fuzzy cognitive maps.
This comparison study is part of our bachelor's thesis under supervision of Sara D'Onofrio (Ph.D. student at the University of Bern (cf. http://www.iwi.unibe.ch/about_us/people/d_onofrio_sara/index_eng.html). She and her research group are investigating the use of fuzzy cognitive maps in the subject area of cognitive computing (cf. https://en.wikipedia.org/wiki/Cognitive_computing). Since fuzzy cognitive maps and graph databases share many characteristics, it seems possible that a graph database is well suited to store fuzzy cognitive maps.
Since we do not want to miss out on possible candidates, we are doing a broad-based study of different graph databases. In the scope of our bachelor's thesis, however, we cannot analyse every graph database as precisely as we would like to. Therefore, we rely on public documentations and user manuals to determine whether or not the graph databases fulfil the basic requirements to store fuzzy cognitive maps.
We hereby wanted to ask you to support our research by having a look at the properties below and provide us with a brief feedback concerning the missing criteria:
Summary: For our study on graph databases for fuzzy cognitive maps storage we would be thankful if you could complement the properties of your graph database listed below.
In the case of TitanDB we found that it allows…
directed edges between nodes
A -→ B
allows nodes to be named
Dog -→ Food
allows edges to be named
Dog --likes→ Food
attributed nodes
Dog(brown, 14kg) --likes→ Food(vegan)
attributed edges
Dog --likes(since 05.02.1998, very)→ Food
many-to-one relationships (Multigraph)
Dog → Food ← Cat
modifiability of once implemented graphs
However, we could not find if TitanDB allows...
the possibility to link/connect two or more databases/graphs to generate a greater knowledge base. For example many database management systems offer the possibility to merge databases in some way. (e.g. export A & B, import to C) We are curious if there is a graph database management system that supports connecting databases/graphs. (e.g. given graph A and B, a connection would result in a knowledge base A---B, with A and B remaining individual databases/graphs)
compatibility with additional web standards of w3c (https://www.w3.org/TR/#w3c_all)
(like RDF, SPARQL, N-Quad, OWL, SKOS, ODBC, JDBC, XML, XPath, JSON, CSV, ...)
We would be very grateful if you could fill our knowledge gap. Complete transparency about your graph databases abilities would allow us to further consider TitanDB as a fuzzy cognitive maps storage possibility and include it in the subsequent and more detailed research.
Thank you in advance for your help. If you have any questions, please do not hesitate to ask us.
Best regards,
Alex Kräuchi & Manuel Grutsch
Marko Rodriguez
(like RDF, SPARQL, N-Quad, OWL, SKOS, ODBC, JDBC, XML, XPath, JSON, CSV, …)
Manuel Grutsch
Thank you very much for your insights and your support for our research. Your answer helped us clarify the remaining uncertainties.
Best regards and thanks again,
Alex & Manuel