Heuristics-based Semantic Annotation for Deep Web Query Results

Deep Web data is encoded in HTML pages through query-processing, the structural information of database mode is completely lost. In addition, current query-results are only for human browsing. For a higher use value of query-results, they must be understandable and processable for machines. Deep Web semantic annotation aims to add correct semantic to query-results, which enables the computer understand and process these data. This paper proposes a heuristics-based semantic annotation method. According to the characteristic analysis of interface page and result page, this paper summarizes some heuristic information. This method uses this heuristic information in turn to analyze the data to be annotated, for identifying a semantic vocabulary for each data unit. Finally, it performs a semantic annotation experiment on the Deep Web data of various areas in the UIUC standard dataset. The experimental result indicates our approach is highly effective. Compared with Ontology-based annotation (OBA) method, our method has a better performance.