The future
Stephen Wolstenholme
Your data is imported into a grid and used to train a SwingNN neural
network. The input values are forced to swing beyond the training
limits. The output values are forecasted. A new neural network is
created using the new inputs and forecasted outputs. The inputs are
adjusted and another neural network is created. The process continues
until a neural network is found that has characteristics close to the
original but with inputs beyond the limits. That neural network is
then used to produce new forecasts. The new forecasts can be used to
produce more forecasts.
Steve
--
Neural network applications, help and support.
Neural Network Software. www.npsl1.com
EasyNN-plus. Neural Networks plus. www.easynn.com
SwingNN. Forecast with Neural Networks. www.swingnn.com
JustNN. Just Neural Networks. www.justnn.com