SISTEMAS INTELIGENTES PARA TEXTOS DA WEB / INTELLIGENT SYSTEM FOR WEB TEXTS
AUTOR(ES)
FERNANDO HIDEO FUKUDA
DATA DE PUBLICAÇÃO
1999
RESUMO
This dissertation investigated the application of KDD´s process and artificial intelligence techniques for the development of a WEB text evaluation tool according to a specific interest. The work, in computing systems area, had as objective the development of techniques together with computing intelligence techniques to automate the texts evaluation by a tool, as well the development of a prototype of this tool. The techniques developed in this dissertation had as objective evaluate taext automatically with independence of the language used and without using NLP (natural Language Processing) techniques. Another onjective was develop this work based on others works as a way to continue the scientific development. The prototype, SITEX - Intelligent System for WEB Texts - was developed by modules which provides the condition to choose the best tool to solve a problem. Another objectiv on the development of the prototype was to use the C++ language, which is largely used and well known by artificial intelligence techniques developers, providing in this way facilities for development and integration of others modules. The techniques was also developed with the objective for WWB standards treatment, as the TCP/IP protocol, lamguages as HTML, DHTML, javascript, Vbscript, and others, which provides the usage of theses techniques in the Internet, intranets, Extranets and locally. The case studies werw based on WEB texts provided by on- line newsletters subscriptions and evaluated according to three types of interests. The interests used in these case studies were: ecoonomy, sports and computing. The results found in this case studies shows the validity of the techniques developed in this work. The results os SITEX compared to the result of TextAnalyst, a well known product in the market for Web text analysis based on artificial intelligence, shows the quality of the techniques discovered in this dissertation.
ASSUNTO(S)
assessment of texts artificial intelligence neural networks web avaliacao de textos internet web inteligencia artificial redes neurais internet
