Improving XML schema matching performance using Prüfer sequences

A novel XML document structure comparison framework based-on sub-tree commonalities and label semantics

“…Note that the issue of integrating structural and semantic similarity evaluation has also been investigated in the contexts of schema matching/integration [3,4,18], as well as ontology mapping/mediation [46,51,52]. Yet, while comparable to tree-based XML documents, schemas and ontologies often underline more intricate graph structures, and thus require graph-based algorithms and heuristics in evaluating similarity, which are out of the scope of this paper (here, we limit our presentation to XML tree-based approaches).…”

“…As for precision, our method tends to yield average levels that are identical to those attained using existing comparison methods, which is underlined in the results corresponding to sets S 2 , S 3 , S 5 , and S 6 . In a few cases, it achieved lower precision, i.e., with sets S 1 and S 4 .…”

“…Section 2.2.2), in comparison with our study (as well as N & J [55] and DCWS [16]), which targets result quality (e.g., clustering effectiveness [16,55]) and higher comparison accuracy. Nonetheless, we are currently investigating several techniques related to XML similarity and performance enhancement, such as Prufer sequence encoding [4], B-Tree indexing [19] and Entropy [31], aiming to improve our method"s performance level, without however affecting its effectiveness and result quality. …”

Combining Schema and Level-Based Matching for Web Service Discovery

“…The approach consists of two stages: Prüfer Sequence Construction and Similarity computation. 3 . The Pre-processing phase is considered with the representation of each abstract item (subtree) as a sequence representation using the Prüfer encoding method.…”

“…Let n be the average operation size, k be the average operation number, and S be the number of input WSs. Following the same process in [3], it can be proven that the overall worst-case time complexity of the schema matching algorithm between two WSs is O(n 2 k 2 ) .…”