Neural Networks - page 11

 
finimej:
Hi,

Hi,

I moved your post to this thread where discussion is on the run.

 

Actually, I start with sibus indicator (see attached picture), using HMA period 14 (type of EMA on closed price), and cross with normal EMA period 21 on closed price, and confirm with RSI >50 or RSI <50 to produce the signals. very simple. KISS.

The only headache is to filter out the signales when the market is doing wigsaw. I have tried Bollingerband, or calculated the slope of the MA, or the distance between signales, or calculate angle of the MA, or using ADX. However, this is not realy help. The only one that helped to filter out the none-wanted trading signals there, is the single layer neural network.

The idear here is to

1) use very simple and robust metod such as EMA cross or MACD signal. methmatical rules, this reduce the number of variables.

2) then use NN to enhance the results that produced by signales above. to check out the profits and loss ratio. and filter out the bad entries. so max 4 variables here.

The system shall stands solely on its own, if we do not use the NN to enhance it.

 

Here I got the sidus indicator attached. and the signals on the picture that I would like to filter out. .source

for sidus indicator is from Forexfactory

Files:
sidus.mq4  8 kb
test.gif  32 kb
 
finimej:
Is there any code that program the process of optimization? so that we can automate the optimization.

logic.

0) do only at weekend.

1) set parameters in these range, 0. 200 and with step 1.

2) get the optimization result

3) round of the result of profits factor, to 1.0 digits, so that 7.4=7 and 7.5 = 8.

4) then select the lest trading number in the catagori of top 2 level of profits factor range, that's optimization result that I want.

5) put the new setting into the expert EA and run for next week.

Can the optimization part be coded?

Maybe this article can help you:

Automated Optimization of a Trading Robot in Real Trading - MQL4 Articles

 

forex_nn new fix version?

Sorry for my limited English. It's a neuro source code. Just download from download link of the original author site below.

http://cortex.snowcron.com/cortex.zip

The fixed version(forex_nn_05b.tsc) seems worked good cause with the new weights. Who can help to convert it to metatrader indic & EA?

Thanks in advance!

 

neural network i think best forcasting in forex, stock or commodt

 

question to above signature haba, how do you using the neuroshell with Metatrader?

Anyone success in using NOXA CSSA in neuroshell and connected it with Metatrader platform?

 

train data source

HiddenOx:
I am starting to make my experiments here. Unlikely I am from a different background. My degree is Artificial Intelligence so more technical towards AI techniques such as NN and more newbie in forex unlike the rest of the guru's in this forum...

using the historical price to Predict the future price it self is kinda impossible due to many factors..

as a sum up from the previous experiences.. whats are the good inputs and there expected outputs for a NN ..

anyways,, whats your opinion about whats the good values/indicators to be used as inputs in Neural Network and what would be there expected output(indicators, curve movements..etc..)..

I will test it on a series and different types of neural networks with back error probagation. also found interesting publication about genetic mutation for the results. I just have some problem now on where to find reliable data to feed the NN... where do you use your historical data .. any sources?

Regards,

Hidden Ox.

you can use mT4 export data to mysql (there is sample code in mq4 code base ),and then train data in matlab (matlab has database toolbox)

 

Neural Network results

Hey I've been working on building a neural network that can accurately predict the GBPJPY on a 1H chart and I think I might have done it. I fed price and indicator data from GBPJPY and other pairs (USDJPY,GBPEUR etc) into a back propagation network and trained over the last ten years data. Here's a few pictures of the target/output plot. The output is the green line and the target is the blue line. I'm not sure about you but an error of 4.2x10^-26 is enough for me = ) I've zoomed in to the most recent data to show just how accurate it seems. Now just to implement that network into a working EA and ill be set anyone got any suggestions?

Files:
gbpjpy60.jpg  37 kb
gbpjpy60-1.jpg  37 kb
gbpjpy60-2.jpg  48 kb
 

You've made some basic mistakes:

- using typical feed forward neural nets for predicting time series like currencies exchange rates is a very poor idea

- you're trying to predict values at H1 time frame - it can't be done for Forex with good results. Use D1 or H4 (for currencies that have low volatility)

- you use to much data as an input - neural net gets "to used" to training data and it will perform very poor in live trading

- you're getting to excited looking at training data

- it is impossible to train a neural net that will work effectively for a long time. Typical neural net trained to predict time series gives about 20-100 good predictions and then it has to be retrained to fit recent changes

If you want to create useful neural nets for time series predictions read about evolutionary neural networks (feed forward neural networks coded as flexible neural trees; their architecture is optimized using PIPE or GEP; flexible activation function parameters are optimized using PSO, EPSO or simulated annealing etc.)