Better NN EA development - page 29
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
With respect, I do understand the indicators workings. And I also agree with what you are saying! My main point was that all the indicators are using weights that were calculated in 2003 - are you telling me that these weightings are as relevant now as they were then?
FWIW, I have registered Cortex, so can now save the NN file. Unfortunately, this is not a simple list of weights as seen in the indicator, in fact the miriad of neuron weights bear little resemblance to what one might expect they 'should' be. (at least I haven't figured it out yet!) The other annoyance is that Cortex itself is limited in the type of 'lag' file configurations it can generate, and the only way to aquire the required flexability is through scripting - but this means that you must pay again to enable this feature! My advice to anyone thinking of registering Cortex - don't, it simpy allows saving the inflexible NN file! Instead, for the same price, register the scripting feature, you then can access all of Cortex's features...
With respect, I do understand the indicators workings. And I also agree with what you are saying! My main point was that all the indicators are using weights that were calculated in 2003 - are you telling me that these weightings are as relevant now as they were then? FWIW, I have registered Cortex, so can now save the NN file. Unfortunately, this is not a simple list of weights as seen in the indicator, in fact the miriad of neuron weights bear little resemblance to what one might expect they 'should' be. (at least I haven't figured it out yet!) The other annoyance is that Cortex itself is limited in the type of 'lag' file configurations it can generate, and the only way to aquire the required flexability is through scripting - but this means that you must pay again to enable this feature! My advice to anyone thinking of registering Cortex - don't, it simpy allows saving the inflexible NN file! Instead, for the same price, register the scripting feature, you then can access all of Cortex's features...
Honestly i do not know if the weights are as efficient in todays market as they were in 2003.. I am assuming they are not. I guess only an update set would tell.. but it is a fact the current set of weights are still able to produce profitable results..
btw where are you registering cortex software?
You 'register' Cortex by buying it.
SnowCron.com order page
You 'register' Cortex by buying it. SnowCron.com order page
There are some many links. Which link you are referring pls.
There are some many links. Which link you are referring pls.
With this link you purchase registration details for Cortex (you click the 'Buy Now' button and pay via Paypal). If you want to buy the Scripting features of Cortex, you would select that option from the Order page instead...
Backpropagation Neural Network Application
I have registered scripting and Cortex, but i haven't good results. First in samples have errors - replace all MVMIN, MVMAX with MV_MIN, MV_MAX.
For me usefull are script 4b, 5a and 5b, but i dont't know how author is generated these weghts.
I have registered scripting and Cortex, but i haven't good results. First in samples have errors - replace all MVMIN, MVMAX with MV_MIN, MV_MAX. For me usefull are script 4b, 5a and 5b, but i dont't know how author is generated these weghts.
Yep, I've hit the same impasse!
Having spent several hours looking for 'bugs', I've reached the conclusion that the 'working' weights were generated with something else, (or some other code) as the code as listed works fine (apart from the error you pointed out) but the weights are a mess! In fact, the more you train the network, the larger individual weights become - something is definitely wrong!
In Cortex have something wrong.
I train with my own backpropagation code this indicator with data from 2007 to today. May be have forecast for one bar.
Some results from tests:
noc_cortex.htm
NOC with Dimicr NN training
noc_dimicrnn.htm