Genetic-Fuzzy Programming Based Linkage Rule Miner (GFPLR-Miner) For Entity Linking In Semantic Web

Tracking #: 1663-2875

This paper is currently under review
Amit Singh
Aditi Sharan

Responsible editor: 
Claudia d'Amato

Submission type: 
Full Paper
Semantic Web Data Sources follow Linked Data principles to facilitate efficient information retrieval and knowledge sharing. These data sources may provide complementary, overlapping or contradicting information. In order to integrate these data sources, we perform Entity Linking. Entity Linking is an important task of identifying and linking entities across data sources that refer to the same real-world entities. In this work, we have proposed a Genetic Fuzzy approach to learn linkage rules for Entity Linking. This method is domain independent, automatic and scalable. Our approach uses fuzzy logic to adapt mutation and crossover rates of Genetic Programming to ensure guided convergence. Our experimental evaluation demonstrates that our approach is competitive and make significant improvements over state of the art methods.
Full PDF Version: 
Under Review