Improving the ShExML engine through a profiling methodology

Submitted by Herminio Garcia... on 03/29/2024 - 03:49

Tracking #: 3680-4894

This paper is currently under review
Authors: 
Herminio Garcia-Gonzalez

Responsible editor: 
Guest Editors KG Construction 2024

Submission type: 
Tool/System Report
Abstract: 
The ShExML language was born as a more user-friendly approach for knowledge graph construction. However, recent studies have highlighted that its companion engine suffers from serious performance issues. Thus, in this paper I undertake the optimisation of the engine by means of a profiling methodology. The improvements are then measured as part of a performance evaluation whose results are statistically analysed. Upon this analysis, the effectiveness of each proposed enhancement is discussed. As a direct result of this work the ShExML engine offers a much more optimised version which can cope better with users' demands.
Full PDF Version: 
PDF icon swj3680.pdf
Tags: 
Under Review
Long-term Stable Link to Resources: 
https://zenodo.org/doi/10.5281/zenodo.10890036