An object-oriented approach on Web information representation and derivation

Abstract

A framework on Web information representation, extraction and derivation utilising the object-oriented approach is proposed. We define a structure which is similar to a class in object-oriented design and programming, and name it as an extended class (eclass). An eclass contains data attributes, member functions/methods, inference rules and presentations, and therefore, facilities information derivation and Web presentation. As an object-oriented approach, the framework also supports encapsulation, polymorphism, composition and inheritance. It can be implemented as an extension of an existing object-oriented programming language. An eclass can be extremely effective in content-dependent entity description. It describes a value under its context, consequently, the search engine not only searches the value, but also understands the meaning of the value in the context. We argue that information in any form or complexity can always be represented by instances of eclasses.