Summary: | 碩士 === 國立中興大學 === 資訊科學系所 === 94 === Abstract
XML is nowadays the standard for electronic data interchange on the Internet. The explosive volume of information made available on the Internet has influenced the de-velopment of a new class of applications based on selective data dissemination, which is a technique of sending data stream selectively. In this thesis, various approaches of bringing together automata theory and the XML data stream are demonstrated and analyzed.
Another focus is on how to query streamed XML documents using a subset of XPath language – Linear XPath. The syntax of Linear XPath is similar to regular expression, and therefore many techniques commonly used in regular expression are applicable to Linear XPath. The algorithms proposed in this paper are related to union operation of deterministic finite automata (DFA). That is, for two automata M and N, how to find an automaton M′ accepting language L(M) ∪ L(N). The problem of union operation for finite automata has been discussed by several articles, but a common drawback among them is the need to reconstruct the entire DFA for every new union operation. The innovation of our work is a DFA union algorithm that does not require the reconstruction of the existing automata. The advantage is clear especially when the number of XPath expressions is large, where the computation time of DFA union operation is effectively reduced.
Our works has shown that the union operation of XPath Deterministic Finite Automata (XDFA) can be done without reconstructing the entire automaton, making it possible for the DFA-based approach to provide high-performance XML filtering for large number of queries that support dynamic Linear XPath expression insertions.
|