| |
Abstract:
We explain how the training data can be separated into a
cleaned part and an unexplainable noisy part. Analogous to the
data, the neural network is separated into a time invariant
structure used for forecasting, and a noisy part. We propose a
unified theory connecting the optimization algorithms for cleaning
and learning together with algorithms that control the data noise
and the parameter noise. The combined algorithm allows a
data-driven local control of the liability of the network
parameters and therefore an improvement in generalization. The
approach is successfully evaluated for the task of forecasting the
German bond market.
|
