Conclusion

Neural networks are suitable for predicting time series mainly because of learning only from examples, without any need to add additional information that may create more confusion than predictive value. Neural networks are able to generalize and are resistant to noise. On the other hand, it is generally not possible to determine exactly what a neural network has learned, and it is also hard to estimate the possible prediction error.

However, neural networks were often successfully used for predicting time series. They are especially useful when we do not have any other description of the observed series.

About

This text and the demonstration were created as a semester project at the Czech Technical University in Prague by Marek Obitko.

ยท