Using data-mining for short-term rainfall forecasting

Abstract

Weather forecasting [12] has been one of the most scientifically and technologically challenging problems around the world in the last century. This is due mainly to two factors: firstly, the great value of forecasting for many human activities; secondly, due to the opportunism created by the various technological advances that are directly related to this concrete research field, like the evolution of computation and the improvement in measurement systems. This paper describes several techniques belonging to the paradigm of artificial intelligence which try to make a short-term forecast of rainfalls (24 hours) over very spatially localized regions. The objective is to compare four different data-mining [1] methods for making a rainfall forecast [7], [10] for the next day using the data from a single weather station measurement. © 2009 Springer Berlin Heidelberg.