Using neural networks in trading - page 17

New comment
 
Figar0:
You make it sound as if you have this hidden layer that has a life of its own) But I think I know what you mean... I'm practicing the opposite method. I find a lot of different solutions, for example on one month's data, then I check it for 2 months, discard the solutions that don't fit the new data, add a lot of solutions, etc.

Hm... That is we show the net 1 combination of inputs and check with one output. We weed out the scales that are inappropriate. Then we show the second inlet. And so on, until we are left with one solution? So this will be the limit that the network can "learn". This is if we go back to the question of retraining.
 
grell:

Hmmm... So we show the network 1 combination of inputs and check against one output. We sift out the scales that are not suitable. Then we show the second intake. And so on, until we are left with one solution? So this will be the limit that the network can "learn". This is if we go back to the question about retraining.
Why overtraining if only solutions capable of working outside the data already seen are selected? Just the opposite, up to a reasonable limit of course.
 
Figar0:
Why overtraining if only solutions capable of working outside the data already seen are selected? Just the opposite, to a reasonable limit of course.

I may have exaggerated:) You know what I mean.
 
grell:
I will post a screenshot of the retrained network soon.

Last year, I was already ready (and prepared the necessary material) to start a topic on the forum called "Natural Intelligence Lab", where it was supposed to take a few months of leisurely research of known theories and practices from the perseptron to multilayer networks (with your help, among others).

It was supposed to look like an institute/university lab work. No theory. Only practical investigations of the different networks. At the beginning of each lab work only references to basic concepts, lectures and articles about the topic under study are included.

Objective: To interest as many forum members as possible in neural networks and to teach them how to build various types of networks themselves.

Main objective: In the end (after one year) to show that neural networks are NOT CAPABLE TO WORK PROFITABILITELY IN THE FINANCIAL MARKET.

And to putan end to this topic in the end.

 
DhP:

Last year, I was already ready (and prepared the necessary material) to start a topic on the forum called "Natural Intelligence Lab", where it was supposed to take a few months of leisurely research of well-known theories and practices from the perseptron to multilayer networks (with your help, among others).

It was supposed to look like an institute/university lab work. No theory. Only practical investigations of the different networks. At the beginning of each lab work only references to basic concepts, lectures and articles about the topic under study are included.

Objective: To interest as many forum members as possible in neural networks and to teach them how to build various types of networks themselves.

Main objective: In the end (after one year) to show that neural networks are NOT CAPABLE TO WORK PROFITABILITELY IN THE FINANCIAL MARKET.

And putan end to the topic at the end of the day.

Have you ever done all this yourself? Because a lecturer should know what he teaches .... It's not like the director of a music conservatory doesn't have to know how to play the violin.
 
DhP:

Last year, I was already ready (and prepared the necessary material) to start a topic on the forum called "Natural Intelligence Lab", where it was supposed to take a few months of leisurely research of well-known theories and practices from the perseptron to multilayer networks (with your help, among others).

It was supposed to look like an institute/university lab work. No theory. Only practical investigations of the different networks. At the beginning of each lab work only references to basic concepts, lectures and articles about the topic under study are included.

Objective: To interest as many forum members as possible in neural networks and to teach them how to build various types of networks themselves.

Main objective: In the end (after one year) to show that neural networks are NOT CAPABLE TO WORK PROFITABILITELY IN THE FINANCIAL MARKET.

And putan end to the topic at the end of the day.

Have you checked it all? That's about 20 years of work. Give up too early:)))
 
DhP:

1. Last year, I was already ready (and prepared the necessary material) to start a topic on the forum called the "Natural Intelligence Lab", where it was supposed to take a few months of leisurely research of known theories and practices from perseptron to multilayer networks (with your help, among others).

It was supposed to look like an institute/university lab work. No theory. Only practical investigations of the different networks. At the beginning of each lab work only references to basic concepts, lectures and articles about the topic under study are included.

Objective: To interest as many forum visitors as possible in neural networks, and to teach them to create various types of networks themselves.

2. Main objective: At the end (in a year) to show that neural networks are NOT CAPABLE TO WORK PROFITABILITELY IN THE FINANCIAL MARKET.

And putan end to this topic at the end of the day.

Knowing item. 2, tasks in paragraph 1 IMHO will not be performed at a desirable level.

Here, it seems, everyone has their own way ...

But, a joint project - would not hurt IMHO,

And then people would have decided on the choice of grid and parameters and everything else, but it's unlikely it will all be transparent ...

After all, people pay money and travel abroad to learn how to use NeuroShell, for example...

 
It's all very trivial. "There are no fish here," shouts one who is already catching them, or one who is desperate to find them. But there are fish, and they catch them, secretly. But nobody shares. Everyone has their own way. The forum shortens it.
 
solar:
Did you do it all yourself? Because a lecturer should know what he's teaching .... It is not as if the director of a music conservatory does not have to know how to play the violin.


Yeah, I spent almost two years on it. Wasn't lazy to read on the subject, study. I created various networks.

I was getting no result at all. This is not to say that it was always negative. Some results inspired further in-depth study and research.

But the realisation that hundreds of billions of dollars had been spent on Artificial Intelligence in the world to no avail ultimately stopped further waste of my energy and time.

It was painful to part with it all.

1...101112131415161718192021222324...40
New comment