Knowledge Graph OLAP: A Multidimensional Model and Query Operations for Contextualized Knowledge Graphs

Tracking #: 2504-3718

Authors: 
Christoph Schuetz
Loris Bozzato
Bernd Neumayr
Michael Schrefl
Luciano Serafini

Responsible editor: 
Harald Sack

Submission type: 
Full Paper
Abstract: 
A knowledge graph (KG) represents real-world entities and their relationships. The represented knowledge is often context-dependent, leading to the construction of contextualized KGs. The multidimensional and hierarchical nature of context invites comparison with the multidimensional OLAP cube model from data analysis. Traditional systems for online analytical processing (OLAP) employ cube models to represent numeric values for further analysis using dedicated query operations. In this paper, along with an adaptation of the OLAP cube model for KGs, we introduce an adaptation of the traditional OLAP query operations for the purposes of performing analysis over KGs. In particular, we decompose the roll-up operation from traditional OLAP into a merge and an abstraction operation. The merge operation corresponds to the selection of knowledge from different contexts whereas abstraction replaces entities with more general entities. The result of such a query is a more abstract, high-level view -- a management summary -- of the knowledge.
Full PDF Version: 
Tags: 
Reviewed

Decision/Status: 
Accept

Solicited Reviews:
Click to Expand/Collapse
Review #1
By Sven Groppe submitted on 30/Jun/2020
Suggestion:
Accept
Review Comment:

Most of the comments of the reviewers have been sufficiently addressed in the revised version. The authors discuss their proposed research direction KG-OLAP in many details and relatively complete. It is a pity that the authors did not work on performance improvements of their prototype, although I can understand that it is just a prototype for showing the functionality. The authors claim that they want to go for a parallel and distributed version of their prototype. From my experience, the performance could already much improved by using proper indexing and other database optimizations for a local engine. However, understanding the implementation as a proof-of-concept study is maybe enough for a first contribution in this area. Experimental comparisons with other engines of different area than KG-OLAP, but implementing the same functionalities remain as future work.

Review #2
By Aidan Hogan submitted on 26/Jul/2020
Suggestion:
Accept
Review Comment:

I am quite satisfied that the authors have addressed my comments. Perhaps they might still reconsider including some elements of the appendix in the paper, though I understand their rationale for choosing not to do this.

In general I think this is an interesting paper on an important topic, and I recommend that it be accepted.

Review #3
By Maribel Acosta submitted on 29/Sep/2020
Suggestion:
Accept
Review Comment:

(Former Reviewer #3)

I would like to thank the authors for their detailed response and for addressing all my comments.

At this point, I just have a very minor remark about the references to the Appendix. Given the length of the Appendix, I would suggest the authors to specify in the text the Appendix section (or page) that is referenced.