Qualidade de Informação na WEB: Um Prognóstico Fuzzy Baseado em Metadados

Web users have to deal with a huge amount of information that is spread out at many sources, particularly, if they have no precise knowledge of their information needs. They constantly retrieve information that is, or may become, outdated, imprecise or invalid, also not forgetting the intentionally wrong, and false or biased information. Considering that users, a priori, have no means to assess the quality of the retrieved information, part of it is, or will become at some point, useless.
This thesis proposes a model, a methodology and an architecture for Web information quality prediction, based on its metadata. A fuzzy logic approach has been adopted, to implement the automated evaluation mechanism of the proposed methodology, due to its ability to deal with diverse concepts and capture humans’ imprecise knowledge.