To all people in this session:
I'm new to this area, I follow the video on http://www.neuralmarkettrends.com/
to create my own stock price prediction package. This session is very helpful for me to clarify some points. The package works, but I only get the accuracy trend a little bit more than 0.6. I'd like to know how well did you do for your prediction? What's the acceptable practical accuracy? What operator do you use for traning and testing the data? What input variables do you use in your prediction? Here are what I used: W-MultilayerPerceptron, input variables: open, close, high, low, volume, the index and index volume, the close price variance compare with previous date, volume variance.
Thank you for your sharing.