| Summary: | Skyline discovery from a hidden Web database can enable a wide variety of innovative applications in Web information integration area. Although there are many limitations such as the finite number of Web accesses one can issue through per-IP-address, no more than k tuples of all matching tuples one top-k query can return, and the restricted Web interface types, Skyline of hidden Web database can still be obtained by using the basic query method. However, the serious problem left is that the query cost is too high. To solve it, this paper puts forward a heuristic algorithm for getting Skyline tuples of a hidden Web database based on the mixture attributes. Specifically, this paper first analyzes intersecting characters of Skyline tuples by parallel coordinate system, and next defines a search decomposition tree for searching heuristically Skyline tuples of a hidden Web database, and then proves that the tree is guaranteed to discover all Skyline tuples. At last, a heuristic method is given for typical Web interface types. Theoretical analysis and experiments demonstrate the effectiveness of the proposed method and the superiority over baseline solutions.
|
|---|