Extracting Fuzzy Gradual Patterns from Property Graphs

Faaiz Shah 1 Arnaud Castelltort 1 Anne Laurent 1
1 FADO - Fuzziness, Alignments, Data & Ontologies
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier
Abstract : A property graph in a NoSQL graph database engine provides an efficient way to manage the data and knowledge due to its native graph-structure storage. A property graph is a labeled directed graph having nodes and relationships with a set of attributes or properties in form of (key:value) pairs. In this work, we aim at mining such graphs in order to extract frequent gradual patterns in the form of "the more/less A1,..., the more/less An" where Ai are information from the graph, should it be from the nodes or from the relationships. In order to retrieve more valuable patterns, we consider fuzzy gradual patterns in the form of "The more/less the A1 is F1,...,the more/less the An is Fn" where Ai are attributes retrieved from the graph nodes or relationships and Fi are fuzzy descriptions. For this purpose, we introduce the definitions of such concepts, the corresponding method for extracting the patterns, and the experiments that we have led on synthetic graphs using a graph generator. We show the results in terms of time and memory consumption.
Complete list of metadatas

Cited literature [14 references]  Display  Hide  Download

https://hal-lirmm.ccsd.cnrs.fr/lirmm-02085780
Contributor : Anne Laurent <>
Submitted on : Monday, October 7, 2019 - 12:22:59 PM
Last modification on : Monday, October 7, 2019 - 2:46:50 PM

File

lirmm-02085780v1.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : lirmm-02085780, version 1

Collections

Citation

Faaiz Shah, Arnaud Castelltort, Anne Laurent. Extracting Fuzzy Gradual Patterns from Property Graphs. FUZZ-IEEE, Jun 2019, New-Orleans, United States. ⟨lirmm-02085780⟩

Share

Metrics

Record views

87

Files downloads

4