Big Data

112 

Big Data: Storage, Sharing, and Security

The purpose of [9] is to crawl over all entity documents in deep web data sources, such

as the documents containing product names and other attributes. Since the authors have all

query logs of Google, the query logs (its format is < query,url

clicked

,times

clicked

>)toward

a target data source are collected and only queries that are clicked for at least two times are

considered. Then the relevant entity names are extracted from the satisfied log queries. The

extraction is based on the Freebase data [37] that provides 22 million entity names. Finally, all

extracted entity names will be sent to the target data ...