نبذة مختصرة : In this deliverable we report on the work carried out for the evaluation and benchmarking of the matching component of the OpenKnowldge kernel. The main goal of the present work has been a extensive evaluation of the approximate Structure-Preserving Semantic Matching (SPSM) algorithm both on a set of synthesized benchmarks, as well as on a set of real world GIS ESRI ArcWeb services. The results demonstrate the robustness and the performance of the SPSM approach on a large number (ca 700.000) of matching tasks. Moreover, the SPSM approach is capable of reproducing similar results to state-of-the-art matchers when only syntactic variation are considered, while outperforming them when semantic variations are applied. The rest of the deliverable is organized as follows: (1) a brief introduction of the goals and main results of the deliverable and (2) the collection of two papers that describe the approach, the related work and discuss the results so far achieved.
No Comments.