 |
QuantShare
2010-08-05 04:52:21
0
|
|
First select the start and end period, then the learning, validation and testing periods. Example: 30% for each one.
Click on Next
Click on "Add" several times to create three inputs.
In the first row, update "WS". Set the field to three.
In the second row, double click on "Symbol Field" and select "Formula". Click on "Select", then type "rsi(14)".
In the third row, click on "Select" then choose "Volume"
Click on Next
In Model Settings, you see we have 5 inputs, one output and one hidden layer with the 5 neurons. You can update the number of Neurons by updating the appropriate numeric input value.
Click on Next
Type your symbol. The symbol data will be used to train the Neural Network model.
Click on Next three times.
Type your model name, and then click on "Finish".
Now, select the Model and click on Train. You can see how the model learns. For example, in the Learning Chart tab, the red dots are the predictions and the black dots are the real values.
The function: Predict("", "Model Name")
Returns the forecast value for each trading bar.
(Note: Please do not respond to the notification emails, instead post your question or response in this forum)
|
|
0
