Automatic construction of vertical search tools for the Deep Web

摘要: With the constant increase in volume of information available on Web, it is more dificult to find specific related a given domain. Users are facing problem overload, which query about specialized subject (local information, e-commerce: hotels, airlines, car rental; science: biology, mathematics, medicine, etc.) web search engine, returns lot pages or results that most cases outside domain interest. This one reason why vertical tools have become necessity for users seek specific-domain from diferent databases Web through input sources called Query Interfaces (ICWs). paper describes an approach automatic integration ICWs, crucial task construct tools. The proposed methodology validated by realizing prototype VSearch allows transparently multiple unified ICW. ICWs based on: i) hierarchical model AEV modeling visual content ICW; ii) semantic clustering identification relationships between fields ICWs; and iii) field homogenization unification process schemes construction was implemented evaluated using study case. experimental demonstrate high precision phase efective create functional tool