DMCA
Annotating structured data of the deep Web (2007)
| Venue: | In: Proc. of the IEEE 23rd Int’l Conf. on Data Engineering. Istanbul: IEEE Computer |
| Citations: | 2 - 0 self |
BibTeX
@INPROCEEDINGS{He07annotatingstructured,
author = {Hai He and Weiyi Meng and Hongkun Zhao and Clement Yu},
title = {Annotating structured data of the deep Web},
booktitle = {In: Proc. of the IEEE 23rd Int’l Conf. on Data Engineering. Istanbul: IEEE Computer},
year = {2007},
pages = {376--385},
publisher = {Society Press}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Abstract. An increasing number of databases have become Web accessible through HTML form-based search interfaces. For each query submitted to such a Web site, data instances retrieved from the underlying database are dynamically encoded into the returned result pages for human browsing. For the encoded data instances to be machine-processable, which is essential for many applications such as deep Web data collection and comparison shopping, they need to be extracted and assigned meaningful labels. In this paper, we present a multi-annotator approach that first aligns the corresponding data instances on sample result pages into groups, then holistically annotates these groups from different aspects, and aggregate the different annotations to predict a final annotation label for each group of data instances. Then an annotation wrapper for the search site is automatically constructed. Our experiments show that the approach is highly effective for the deep Web annotation. 1
Keyphrases
deep web data instance different aspect corresponding data instance encoded data instance human browsing deep web annotation different annotation web site html form-based search interface result page sample result page meaningful label many application search site deep web data collection annotation wrapper comparison shopping final annotation label underlying database multi-annotator approach
