Neural nets are useful when you have some idea of some predictive indicators or other time series, but have no idea how to put them together into a set of rules for trading. Neural nets make predictions (the output) about the future value of some data series based on some indicators that you feed into them (inputs). The neural net uses some historical data (the training set) to “learn” how to make accurate predictions of the output using the inputs.
The neural network learns the relationship between inputs and the time series you want to predict based on historical data. Once the model is trained, it may be applied to new inputs and produce a predicted output.
If the prediction is greater than a threshold, such as when the predicted percent change in open is greater than 0.3 in this model, a buy signal is generated. If the prediction is less than a predefined threshold value, a sell signal is generated.