Abstract:The Semantic Web opens up new opportunities for the data mining research. Semantic Web data is usually represented in the RDF triple format (subject, predicate, object). Large RDF-style Knowledge Bases contain hundreds of millions of RDF triples that represent knowledge in a machine-understandable format. Association rule mining is one of the most effective techniques for detecting frequent patterns. In the context of Semantic Web data mining, most existing methods rely on users intervention that is time-consuming and error-prone due to a large amount of data. Meanwhile, rule quality factors (e.g. support and confidence) usually consider knowledge at the instance-level. Namely, these factors disregard the knowledge embedded at the schema-level. In this paper, we demonstrate that ignoring knowledge encoded at the schema-level negatively impacts the interpretation of discovered rules. We introduce an approach called SWARM (Semantic Web Association Rule Mining) that automatically mines Semantic Association Rules from RDF data. The main achievement of SWARM is to reveal common behavioural patterns associated with knowledge at the instance-level and schema-level. We discuss how to utilize knowledge encoded at the schema-level to add more semantics to the rules. We compare the semantic of rules discovered by SWRAM with one of the latest approaches in this field to show the importance of considering schema-level knowledge. Initial experiments performed on RDF-style Knowledge Bases demonstrate the effectiveness of the proposed approach.

An Approach for Discovering Frequent Patterns from Log Ontologies with DL-safe Rules

Evaluation of Access Patterns from Log Ontology with DL-safe Rules

An ILP Approach for Mining User Access Patterns Form Hybrid Reasoning System

A Level-Wise Search Approach to Frequent Webaccess Pattern Discovery with Hybrid Reasoning

An Ilp Approach To Mine The Association Rules On Log Ontology

Domain relation learning from events in log ontologies based on hierarchy association rules

ILP Approach of Extending AL-log to Mine Log Ontology

The Datalog(Dl) Combination Of Deduction Rules And Description Logics

Association Rules Mining Based on the Discriminative Concept Lattice

DatalogDL: Datalog Rules Parameterized by Description Logics

Ontology Matching Based On Probabilistic Description Logic

Research on Parallel Frequent Pattern Discovery Based on Ontology and Rules

Discovering Sequence Association Rules with User Access Transaction Grammars

Research on parallel frequent pattern mining based on ontology and rules

Rule Extraction in Combining Description Logics with DL-safe Rules

Sequence Association Rule Discovery Based on User Access Transaction Grammar

Mining semantic association rules from RDF data

Semantic Text Mining with Linked Data

A Taxonomic Relationship Learning Approach for Log Ontology Content Event.

Exploring Logical Rules Based on Causal Semantics Analysis of Relational Data

An effective rule miner for instance matching in a web of data.