MultiAddIntent: Efficient RDF Data-Oriented Incremental Construction Concept Lattice Algorithm

Liu Yang, Guohui Li, Meihong Xiao, Shuai Luo, Yangying Tan, Ziqiang Tang, Shuai Mu

Abstract: Formal concept analysis (FCA) is an effective approach for data mining and knowledge representation. Inspired by AddIntent algorithm for ordinary concept lattice construction, we propose a novel incremental construction algorithm Mul-tiAddIntent for building RDF data in semantic web based on FCA. We present a lattice-based schema by pattern structures to deal with multi-attribute values for RDF data. A hash table is built to record intent intersections to avoid recursive calls and partial order calculations during the incremental construction, so it can speed up concept lattice construction of RDF data. We utilize mathematical methods to prove the accuracy of our MultiAddIntent algorithm. The experimental results on random datasets with different densities show that our MultiAddIntent algorithm achieves significantly faster construction time than the algorithms of AddIntent and FastAddIntent, and the experimental results on DBLP show that our algorithm outperforms the two other algorithms, expecially on the large and dense datasets, so our algorithm is more scalable in real applications.