Fog Formation Prediction In Coastal RegionsUsing Data Mining Techniques

The recent growth of many scientific and business databases created a need of a new class of tools and techniques for automated search the knowledge hidden in these databases. For this reason it became necessary to integrate data mining and database technology on scalable parallel platforms. This paper discusses a parallel implementation of a data mining strategy developed for MIMD platforms: objected related data mining technology, that enables a specific form of data mining optimized for better performance and more complete results. Our application is related to observed meteorological conditions at the International Airport of Rio de Janeiro. The main objectives are to understand the parameters that have influence on visibility/weather conditions (fog formations), and to compare with current methodologies to predict future situations with some antecedence. In our analysis we have considered the last 20 years to build a model and extract the knowledge to make predictions. Transactions on Ecology and the Environment vol 18, © 1998 WIT Press, www.witpress.com, ISSN 1743-3541 166 Environmental Coastal Regions