Pattern discovery in semantic data mining

Date: February 6, 2015

Speaker: Agnieszka Ławrynowicz, Poznan University of Technology, Poland

Abstract

In this talk, I will describe an algorithm, named Fr-ONT-Qu, for mining patterns from Semantic Web data. I will discuss the application of Fr-ONT-Qu in two use cases.

Firstly, I will talk about how Fr-ONT-Qu might be used for pattern based classification of semantic data, and I will present the experimental results. This will include a discussion of the application of Fr-ONT-Qu to meta-mining, where I will describe how Fr-ONT-Qu was used inside RapidMiner to mine knowledge discovery (KD) workflows with use of the Data Mining OPtimizaion Ontology (DMOP).

Secondly, I will describe our ongoing work on ontology learning from Linked Data, where Fr-ONT-Qu is being adapted for the task of pattern based ontology axiom induction. This will also include a discussion on how we plan to re-use Fr-ONT-Qu inside a Protege plugin being developed for ontology learning from Linked Data.

References

• Pattern based feature construction in semantic data mining. Agnieszka Ławrynowicz and Jedrzej Potoniec., Int. J. Semantic Web Inf. Syst., 10(1) 2014., http://knoesis.org/resources/ijswis/209.pdf
• Data Mining Optimization Ontology. C. Maria Keet, Agnieszka Ławrynowicz, Claudia d’Amato, Alexandros Kalousis, Phong Nguyen, Raul Palma, Robert Stevens, and Melanie Hilario, Journal of Web Semantics, accepted, PDF

Presentation: Slides are available here.