fbpx

Quant development: Neural Networks in Swing Trading

(Last Updated On: July 21, 2011)

Quant development: Neural Networks in Swing Trading

Mixing Back propagation NN algorithms with complementary models in trading strategy design. Technical discuss. White papers are welcomed.

—-

First of all, does there really exist NN based trading models?

—-

I’m also interested by feedbacks about neural network in trading strategy. There are a lot of possibilities using NN.How to train the system ? What kind of Inputs ? Output expected…As far as I’m concerned, I tried in the past to use genetic algorithm for the training.I used as input some technical Indicators ( for instance MMA 50, MMA 10 …).I was expecting to find the direction to take, so my output was a double between -1 and 1. if output at a given time is close to 1(-1), It means that I have to take a long(short) position.My results was quite bad (sorry I don’t remember exact figures)… I didn’t try harder because of busy schedule. Now, I think I will take the time to search again 🙂

16 days ago

—–

there are successful trading models built using NNs. Here is one I built about 3 month ago and it is still outperforms the market and produces results that are consistent with backtested data.

http://www.bowgett.com/TradingSystems/Details.aspx?id=LSM900.Columbia

—-

that’s interesting. We can have further discussion about this. NN totally falls behinds the mainstream algorithms of machine learning. Since 2006, Hinton proposed the concept of deep learning, which is essentially NN with unsupervised pre-training algorithms.

—-

You don’t have to know where stock is going to move to make money. There are option strategies that you can use to make money if stock moves in any direction

—–

Yes, there are successful ANN models. ANN is just a tool to do non-linear regression. SVM is a good alternative. Sample configuration: price time series (multimple time frames) + custom analytics, custom stabilisation filters -> partially recurrent backprop NN (one hidden layer, bipolar sigmoind transfer function) ANN -> signal processor -> risk management engine -> trading strategy -> market. No point in predicting the price. 90% of the success is in the risk management, imagination and attitude. It’s a minority game.

—–

If you understand back propagation, that is optimisation using gradient descent, then there are very many financial problems you can solve that don’t have analytic solutions. The big problem with classical neural networks (i.e. standard backprop as when I did my PhD on it 20+ years ago) is that they are not very good at regularisation, that is they are prone to overfitting. Some application areas are relatively tolerant to this as there is lots of data, but financial engineering is a game of one-upmanship and you’ve got to squeeze the last bit out of the data you’ve got. So you really need to use Bayesian methods. I recommend Chris Bishop’s latest book, Pattern Recognition and Machine Learning http://research.microsoft.com/en-us/um/people/cmbishop/prml/, but also look at the recent work of one of the founding fathers of backprop, Geoff Hinton,http://www.cs.toronto.edu/~hinton/, he is still forging ahead when most others have dropped by the wayside.

 

 

NOTE I now post my TRADING ALERTS into my personal FACEBOOK ACCOUNT and TWITTER. Don't worry as I don't post stupid cat videos or what I eat!
Don't miss out!
Get my TRADE LIKE A BOSS 2 PDF Books

You will received instantly the download links.

Invalid email address
Give it a try. You can unsubscribe at any time.

NOTE!

Check NEW site on stock forex and ETF analysis and automation

Scroll to Top