Training NeuralCode with Stock information for predictionNext, let’s try to use NeuralCode for Stock Trading.
- Launch the Predict Stock Prices Worksheet.
- Click on "NeuralCode" to launch the training and prediction dialog.
Click on the "Input Range" text box.
Select cells A2 to D11. Make sure you see 'Predict Stock Prices'!$A$2:$D$11 in the text box.
Click on the "Target Column" text box.
Select column "F" in the text box.
We are training the networks with Tomorrow’s Closing price to give it the ability to predict future prices with historical information.
Click on the "New Training" button.
Wait till you see the message "Training Completed Successfully". This message will appear when the Error Rate falls below the Target Error Rate.
Without closing the dialog, click on the "Predict Results" tab. Click on the "Input Attributes" text box.
Select cells A12 to D12. Make sure you see 'Predict Stock Prices'!$A$12:$D$12 in the text box.
We are predicting tomorrow’s closing price with the historical data from A2 to D11 in the Neural Networks and today's Open, High and Low.
- Click on the "Output Results" text box. Select column "H".
- Click on the "Predict" button and you will see the results.
Train the Neural Networks with the different Technical Indicators that you are accustomed with.
Setting a smaller Target Error Rate will allow you to have more accurate results. However, more training time will be required.
It is not necessary that you see "Training Completed Successfully" to perform the prediction. You can click on "Stop Training" and proceed to the Prediction. Many industry experts have warned of "over training"/"over fitting" in neural networks and missing the big picture.