作者: Georg Lausen , Thomas Hornung , Kai Simon
DOI:
关键词:
摘要: Deep Web (DW) sources offer a wealth of structured, high-quality data, which is hidden behind human-centric user interfaces. Mashups, the combination data from different services with formally defined query interfaces (QIs), are very popular today. If it would be possible to use DW as QIs, whole new set feasible. We present in this paper framework that enables non-expert users convert into machineprocessable QIs. In next step these QIs can used build mashup graph, where each vertex represents QI and edges organize flow between To reduce modeling time increase likelihood meaningful combinations, assisted by recommendation function during time. Finally, an execution strategy proposed queries most likely value combinations for