Why Neural Nets are Useful for Traders
The difficult classic investment decision problem is consolidating multiple intermarket factors into a clear buy/sell decision. The basic problem is that many traders have several favorite indicators but have great difficulty making a decision because some indicators say "buy" while others may indicate "sell". Multiple indicators are used because no one indicator has enough reliability for confident decisions. The problem for the trader is to make an intuitive decision based on the available, often contradictory, data. Unfortunately making intuitive decisions under stress (due to fear of loss) is the worst circumstance for good decisions. Have you ever made poor decisions due to stress and fear?
One way to deal with this circumstance is to employ an objective mathematical decision making process not affected by human emotion. Consolidating multiple information sources into a clear buy/sell decision is not straightforward for most traders and can require a relatively high level of mathematical skill and much effort to define an applicable model. This is where neural networks can be very useful. They have the ability to, in effect, generate their own mathematical model, based on historical examples, thereby allowing useful decision making while requiring only a small fraction of the time and effort of other modeling approaches.
The neural net technique is employed to automatically learn the relationship between a set of selected indicators and probable future market behavior. It does this by automatically learning the historical relationship between the selected indicators and subsequent market behavior. The result is a mathematical algorithm which consolidates the indicators into a single clear buy/sell signal. A unique and powerful feature of this process is that indications of buy and sell are early and their execution is delayed! For instance a signal might be generated today but the trade won't be executed until a week from today! This is exactly the opposite of most technical indicators whose signals are late because the indicator employs smoothing to eliminate whipsaw trades. You can see that neural networks can have very significant advantages for the trader.