A Query Language Used on Web for Complex Polymer System Search

摘要

In this paper we propose WQFL (Web Query Formulating Language) - a markup language based on XML (Extensible Markup Laingnage) to formulate chemical structure queries and to search hypermedia information, using, the Artificial Inteligence (Al) different approaches. The search activity use a document structure based method and it can be divided into several parts. In the first part, we search the information concerning polymer structure on Web using a traditional search engine (e.g. AltaVista or Google) and we store first significant found pages. The second part implies the evaluate how well a candidate molecule satisfies the desired target properties and constraints, one needs a fitness function that returns a single numerical "fitness" or "figure of merit". The third part consists in the encoding of the Web pages structural information (the position, the occurrences and the context of some HTML elements), building a matrix. The elements of this matrix will be encoding by means of a special operator to obtain an integer number. In the last part, we can apply AI methods to obtain the most relevant Web page, such as genetic algorithms or a self-organizing feature maps neural network based on the competitive learning. The WQFL language can be processed by a SAX (Simple API for XML) library available on Linux, We intend to implement the proposed framework under Limux environment, using a Web interface.