Aggregating and Semanticizing Unstructured Online Event Information

Nightbat是一家Web 2.0公司，专门从事当地活动和庆典的在线销售门票。本地活动组织者可以通过Nightbat管理公告、注册和日程安排。目前，通过Nightbat提供的服务对活动组织者是免费的，对参与者只收取固定的注册费。Nightbat的成功在很大程度上依赖于其网站上举办的活动数量和每天访问网站的人数，而这现在取决于通过不同渠道进行的营销。最近，Nightbat开始对开发智能网络爬虫技术感兴趣，该技术将自动浏览网站，识别它们是否包含本地事件信息，以提取此类信息添加到其事件信息库中。这种方法的优点是，Nightbat将能够为用户提供一站式的本地活动，这些活动已经通过网络从各种来源自动识别和配置。Nightbat面临的主要挑战是：i）事件信息目前没有集中收集在特定的网站上，它们分散在网络上的各个位置。一些组织者可能只在他们自己的网站上做广告，而另一些组织者可能在社交网络或在线分类广告网站上宣布他们的活动;因此，找到正确的地方来自动提取活动信息是一个挑战; ii）一旦在网站上识别出活动，主要的挑战是从组织者发布的非结构化活动信息中提取结构化信息。由于没有在线发布事件信息的标准，事件组织者通常以HTML格式发布非结构化文本数据，这些数据对于人类来说是可以理解的，但机器无法解释。因此，Nightbat要想有效地从Web上聚合事件信息，首先需要开发智能爬行技术，使其能够在线查找与本地事件相关的信息，其次才能从非结构化事件信息中提取有用的事件信息。

Nightbat is a Web 2.0 company that specializes in the area of online sale of tickets for local events and gather-ings. Local event organizers can manage announcement, registration and scheduling through Nightbat. Currently the services provided through Nightbat are free for event organizers and only a fix per registration charge is made for the participants. The success of Nightbat relies heavily on the number of events that are hosted on its website and the number of people who visit the website on a daily basis, which is now dependent on the mar-keting that happens through different channels. More recently, Nightbat has become interested in developing intelligent Web crawler technology that would automatically browse Websites, identify whether they contain local event information to extract such information to be added to their event information base. The advantage of this approach is that Nightbat will be able to offer its users a one-stop shop for local events that have been automatically identified and collocated from various sources over the Web. The main challenges that Nightbat faces are: i) event information are not centrally gathered in specific websites at the moment and they are dis-persed over various locations across the Web. Some organizers might only advertise on their own Website, while others might announce their event on social networks or online classified advertisement websites; there-fore, finding the right places to automatically extract event information is a challenge; ii) Once an event is identi-fied on a Website, the main challenge is to extract structured information from the unstructured event infor-mation that has been posted by the organizers. Given there are no standards for publishing event information online, event organizers publish their events as unstructured textual data often in HTML format that is under-standable for humans but not interpretable by machines. Therefore, for Nightbat to be able to effectively aggre-gate event information from the Web, it will need to first develop intelligent crawling technology that enables it to find information related to local events online and second be able to semantically extract useful event infor-mation from the otherwise unstructured event information.