A Continuous Spatial Query Processing for Push-Based Geographic Web Search Systems

Abstract

With the explosive growth of the web, people often need to monitor fresh information about their areas of interest by browsing the same sites repeatedly. Especially, even for a local area, so many pages are created every day that might be a greatly helpful knowledge in daily decision supports. In order to reduce such monitoring work and not to miss chances to meet critical information for a local area, we are developing a Continuous Geographic Web Search System with Push-based web monitoring services. This paper will describe the system architecture to deal with multiple user queries to represent users' geographic attention over multiple data pages incoming from geographic web crawlers. If a newly incoming data page is relevant to a preregistered query, the user who previously registered the query will be informed spontaneously about the new information by our notification service. This paper especially focuses on the problem of how to match multiple data pages and multiple queries as quickly as possible. For the purpose, we will propose a fast matching algorithm based on a spatial join and show a primitive experimental result with synthesis data.