Weight Measurement for Node vs. Arc; and the Extraction of Meaningful Sub-Graph

초록

Social Network based Searches can be applied to various object domains that can be structured by a graph or a tree such as relational DB, XML, Web, bioinformatics and Genetics, RDF/RDFS, social network, streamlining data graphs, etc. Graph approaches have been popular for the IR community and have recently gained comprehensive motivation in the DB community. The initiation of bountiful data sources has accessible an unrelenting requirement for more automatic yet precise data integration techniques. The unadventurous semantics for processing a user keyword query has been so far to find a set of Top-k results listed by documents such that each and every document has to contain all the keywords that the user requests. Recently, this method has been extended to find a set of networked documents, each of which contains any of the query keywords disseminated across these pages but connected by hyperlinks. For the data retrieved from querying multiple sources, most works focus on how to extract the embedded structured data more precisely that is referred to the graph query. In this talk, at first innovative ranking measures to calculate the Web arc weights will be introduced to determine the significance of Web objects including a Network Ranking Measure and a Context Specific PageRank Algorithm. Describing some essential related works for the graph queries, an optimal and top-k Dynamic Steiner-tree algorithms from Web structuring model are addressed. Then the applications for the structural approaches will be introduced such as Web search engine, and mobile web, etc.