作者: Zhuo Zhang , Juan Du , Liming Wang
DOI: 10.1007/S10844-013-0242-Y
关键词: Pruning (decision trees) 、 Cardinality (SQL statements) 、 Computer science 、 View 、 Database 、 Query optimization 、 Web query classification 、 Web search query 、 Formal concept analysis 、 Data extraction 、 Theoretical computer science 、 Data mining
摘要: Few studies have addressed the problem of extracting data from a limited deep web database. We apply formal concept analysis to this and propose novel algorithm called EdaliwdbFCA. Before query Y is sent, analyzes local context K L , which consists latest extracted data, predicts size results according cardinality extent X (X,Y) derived . Thus, it can be determined in advance if or not. Candidate concepts are dynamically generated lower cover current (X,Y). Therefore, method avoids building concrete lattices during extraction. Moreover, two pruning rules adopted reduce redundant queries. Experiments on controlled sets real applications were performed. The confirm that theories correct effectively applied world.