One of the examples we used to show at our training seminars was a neural net with perhaps 15 to 20 columns of random numbers predicting another column of random numbers. The example worked very well. The purpose of the example was to teach methods to prevent overfitting.
Keep the number of inputs low (from 5 to 10 inputs)
Make sure training sets are large enough to cover up, down, and sideways markets, but not so large as to include obsolete patterns
Don’t use too many hidden neurons
Make sure your inputs are detrended
Optimizing over all chart pages can help.
The best method to judge the success of a neural net is to look at performance in out-of-sample data. Are the trading signals hitting peaks and valleys. Is the profit based on one or two “lucky” trades or does it cover multiple trades.
Neural network isnt’ about finding meaningful relationships between real indicator and data.If you look at the Ward newsletters ,their most successful models created by rsi,cci,stochastic indicators which are basically normalized-detrended data.So it is all about the data.How to fit the data.If you want to find descriptive relationships between your indicators and the data try decision trees but without guaranties.
You could look at the contribution factors that are listed for the inputs on the Detailed Analysis section of the Prediction wizard to learn which data streams are more meaningful in making the prediction.
