Evolutionary and Neural Computation for Time Series Prediction Minisite

Advanced techniques

Advanced techniques for time series prediction

Many different methods can be used to deal with nonlinear time series prediction problems. Among them, artificial neural networks [Kou95] and evolutionary computation techniques occupy an important place being able to adequately model the nonlinearity and nonstationarity. Those techniques have the advantage that they do not need to specify a mathematical model describing the behaviour the time series.

Currently, there exists a wide variety of works, in which Advanced techniques of artificial intelligence have been used or combined with the purpose of time series prediction. Some of this works can be found, for instance in [Koz98] [Hit99] [Nas99] [May99] [Zha00] [val03].

In this site, the use of different techniques to tackle the time series prediction problem are briefly presented.

2005, University CARLOS III of Madrid