An agent-based system framework for multi-slot Web information extraction

At present, the scale and diversity of Web information are immense. Acquiring Web information simply relies on search engine which is increasingly unable to meet user needs, thus Web information extraction (WebIE) technology attracts widely attentions. In this paper, a framework of distributed multi-slot WebIE system based on agent is proposed. It includes user agent, mediator agent, wrapper agent, data store agent, page preprocessing agent and corresponding knowledge base. The agents communicate each other and cooperate together to carry out the general goal of the system. Moreover, aiming at multi-slot extraction, the approaches of extraction rule learning and repair are presented, which enable to enhance adaptability of the system.