Novel exploitation of neural network methods in financial markets

The exploitation of neural network techniques in different types of financial market problems is discussed and illustrated by snatching the inherent qualities of networks to problem domains in finance. In particular the author considers the construction of a portfolio under constraints as an optimisation problem amenable to solution by an analog neural network, and short term equities prediction as a problem in nonlinear multichannel time series forecasting, which can be addressed by feedforward networks