Semantic Workflow Retrieval Based on Multiple Features
Abstract— Semantic similarity retrieval workflow is the primary task of semantic workflow reuse. Semantic workflow retrieval methods only focus on structural features similar to existing, ignore the behavior characteristics, affect the retrieved similar semantic workflow in order to improve the overall quality of the work flow, semantic reuse cost, multiple workflow semantic similarity retrieval algorithm is proposed combined with behavior the structure and use of task execution behavior. Close relationship set expression semantics of workflow, combined with the domain knowledge base to construct the semantic workflow task tree index and data index. According to the semantic query workflow, first task relation between neighboring tree index and index data are filtered to obtain the candidate set based on Semantic workflow; then use graph matching similarity algorithm to verify the candidate semantic workflow set, get sorted when the semantic workflow ensemble is selected. Experimental results show that compared with the mainstream semantic workflow retrieval algorithm, the retrieval performance of this method is greatly improved, and the efficiency of workflow reuse is improved.
Index Terms— TAR, index, workflow, semantic.
Rujuan Wang, Lei Wang, Chunling Xu
College of Humanities & Sciences of Northeast Normal University, CHINA
Cite: Rujuan Wang, Lei Wang, Chunling Xu, "Semantic Workflow Retrieval Based on Multiple Features," Proceedings of 2018 the 8th International Workshop on Computer Science and Engineering, pp. 769-774, Bangkok, 28-30 June, 2018.