TISE: A Temporal Search Engine for Web Contents

In this paper, we present a temporal search engine supporting content time retrieval for Web pages, which is called TISE. The main purpose of TISE is to support the Web search on temporal information embedded in Web pages. Compared with commercial search engines such as Google and Baidu, and other temporal search prototypes, which mainly focus on the creation or crawled time of Web pages, our system concentrates on the extraction and search on content time of Web pages, and can provide more meaningful time-based search facilities, such as temporal relation query. In detail, TISE is based on a unified temporal ontology of Web pages, in which different types of time are defined. We introduce a new type of time ldquoprimary timerdquo to denote the most appropriate time describing the content of a Web page. After an overview of the general features of TISE, we discuss the architecture of TISE and some key modules. And finally, experiment results and analysis is conducted based on fix types of temporal-text queries on TISE and www.baidu.com. The experiment results show that TISE is more efficient when processing temporal-text Web queries.