Artificial intelligence - path smoothing

To add comments, please log in or register
Jox90
686
Jox90  

Hi, I found these videos about path smoothing using artificial intelligence. I obviously thought of using it to smooth price, but I don't know if an indicator could be built using this method and if it would repaint or not. An opinion of a good coder would be great... The code used is showed in the second video.

Here are the links:

https://www.youtube.com/watch?v=v0-OUApP_5Q

https://www.youtube.com/watch?v=ffLPf8kI6qg

Sergey Golubev
Moderator
106122
Sergey Golubev  

I found old threads - 

  • Artificial Intelligence EAs threads: How to "teach" and to use the AI ("neuron") EA thread and Artificial Intelligence thread 

CodeBase

Articles

  • more on the search here

Marco vd Heijden
Moderator
8173
Marco vd Heijden  

An indicator is already smoothing the price for example a moving average is an averaged price.

In training you want to get as close to the real data as possible.

This usually means RAW price feed.

If you swap that with an averaged and lagging price feed, from an indicator, the end results are most likely to become less exciting.

But you never know up front so for these things the rule is to try and see what it does.

It would not be the first time that such experiment would present something unusual and interesting.

So for a real answer on your question you would have to run the tests.

Mladen Rakic
106144
Mladen Rakic  
Jox90:

Hi, I found these videos about path smoothing using artificial intelligence. I obviously thought of using it to smooth price, but I don't know if an indicator could be built using this method and if it would repaint or not. An opinion of a good coder would be great... The code used is showed in the second video.

Here are the links:

https://www.youtube.com/watch?v=v0-OUApP_5Q

https://www.youtube.com/watch?v=ffLPf8kI6qg

What he is using is simple ema smoothing (check the end of first video)
Ivan Negreshniy
17496
Ivan Negreshniy  

A good coder uses artificial intelligence for modeling the intelligence of the trader, not for modeling intelligence moving average.

Marco vd Heijden
Moderator
8173
Marco vd Heijden  

You would have to define the trader.

The Artificial intelligence becomes the trader when it starts with nothing and, by training learns to get to the highest reward possible.

I am not sure if the final 'artificial intelligence trader' would trade in the same manner as regular traders.

The consistency in it's decision making is probably not comparable to the consistency of human decision making. 

Neuronets lack the emotional layer.

Ivan Negreshniy
17496
Ivan Negreshniy  
Marco vd Heijden:

You would have to define the trader.

The Artificial intelligence becomes the trader when it starts with nothing and, by training learns to get to the highest reward possible.

I am not sure if the final 'artificial intelligence trader' would trade in the same manner as regular traders.

The consistency in it's decision making is probably not comparable to the consistency of human decision making. 

Neuronets lack the emotional layer.

emotional layer to reproduce human vices such as greed, envy, meanness, etc?
Jox90
686
Jox90  
Mladen Rakic:
What he is using is simple ema smoothing (check the end of first video)
That means that it worths not so much...thank you Mladen, as always.
To add comments, please log in or register