DistSim - Scalable Distributed in-Memory Semantic Similarity Estimation for RDF Knowledge Graphs
2021 IEEE 15th International Conference on Semantic Computing (ICSC), Carsten Felix Draschner, Jens Lehmann and Hajira Jabeen
TL;DR ⏱️
- Scalable Similarity Estimation for RDF Knowledge Graphs
- MinHash with locality sensitivity hashing
- SANSA stack integration
DistSim includes several feature-based graph entity similarity estimators. The Similarity Estimation Pipeline was created by merging generic software modules. DistSim introduces MinHash with locality sensitivity hashing for large-scale RDF data to improve scalability over all-pair similarity estimations. DistSim's modules can be configured using a variety of (hyper-)parameters to manage the trade-off between information considered and processing time. Furthermore, the similarity estimation pipeline's output is native RDF. DistSim is part of the SANSA stack, with documentation in Scala-docs and unit tests.