Machine learning in trading: theory, models, practice and algo-trading - page 2420
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
Wrong end of the line. This is where things of interest are discussed, including those related to trading. The standard TS are stable one in a thousand or more, relatively few deal with neural networks. So it's a survivor's mistake to claim that MO doesn't work. If this thread didn't exist, there would definitely be nothing to read on this forum.
Maxim, do you know the expression "monkey labor"?
It's about the branch.
And by the way, I have monkeys for fun and as fishing rods for poachers, I do not work with them.
You went to the wrong end of the line. This is where the interesting stuff is discussed, including trading related stuff. The standard TS are stable one in a thousand or more, relatively few deal with neural networks. So it's a survivor's mistake to claim that MO doesn't work. If this thread didn't exist, there would definitely be nothing to read on this forum.
Your Expert Advisor is definitely not working on a neural network. You have written it and have not done at least a two or three month forward test to make sure of the EA stability, and you ran to write a colorful advertising text, to trick people for money.
In order to ensure that the market is free of bias and unfair competition, the brokerage firm has been offering a wide range of services on a daily basis.
And Renat had nothing to do with it, as you say, "monkeys" he sells. Good for him, but he does not deceive people!
You would be ashamed to self-remove from the forum, but no, you continue to write clever texts and teach people.
Before criticizing someone, show your works!
Judging by such posts, you are too far from Maksim's level, since you have failed to learn to use those methods from his articles.
Before you criticize someone, show me your work!
Judging by such posts, you are very far from Maksim's level, since you couldn't learn to use those methods from his articles.
You are absolutely right, I am very far from Maksim's level! My conscience won't let me deceive people!
You are absolutely right, I am very far from Maxim! My conscience won't let me cheat people!
Write in the lawsuit that you believed the man, you read the articles, you dreamed about passive income. And he deceived you.
It is a pity they do not put him on the stake now or at least cut off his hand.
We will support you. We are also victims, we trusted him, we crammed a lot of code, even a decent indicator came out of the neural network, and there is no passive income.
There is a desire to do something like training with an "interactive teacher" (I coined the term myself)
The idea is not to give ready-made labels of AMO, but to teach it by itself from a chart by pointing the mouse "here we have to buy by order, and here to close the market", you can also immediately teach market conditions, draw the signs (trend lines, etc.), the algorithm that has learned in the future must reproduce it all by itself ...
Why so? Having read a little bit about psychology and about AI, I came to the conclusion, that a human cannot express what he wants verbally (or in code), he can only show it ...
A professional surgeon can not describe in words how to make an operation, he can only show and control the process of training the student, in the same way a successful trader can not explain in words how to make money, or rather he can tell you in words, but the description is always not accurate and the effect is very little...
I have come to the conclusion that we should not zigzag labels but teach the robot as a child...
There are questions:
1) Since the robot will be constantly learning neuronics probably won't work, the question is how to implement "memory" to the robot so it will get smarter in the future without getting dumber in the past, and I could constantly teach it... I have only one way of implementing "memory" Memory can be done in the form of "rules" and accumulate them in the course of training the robot, the sum of the rules will be something like a robot's knowledge base... Does anyone have any other suggestions about "memory"?
2) If we accept the first option with memory, how can we implement "punishment" for the robot to understand that it is doing something wrong?
super-task for MO and neural networks, to say "STOP" to a trading robot...
Any robot is obtuse and deterministic, it can be trained by it. Identify its area of "definition" at least so it won't jump out
If you just say "no more", that's half the grail, and a big one at that.
We should sue! This is outrageous!
Write in the lawsuit that you believed the man, read the articles, dreamed of passive income. And he deceived you.
It is a pity they do not put him on a stake or at least cut off his hand.
We will support you. We are also victims, we trusted him, we crammed a lot of code, even a decent indicator came out of a neural network, and still no passive income.
super-task for MO and neural networks, to say "STOP" to a trading robot...
What's the difference between "game \stop game" and "open \not open" or "buy \not buy"?
I don't think so, it's a common classification...
The TS is a filter that filters the incoming market data and gives an output of buy/not buy.
A neural network is also a filter that gives a yes/no output
A neural network and TS are the same filter, except the NS filter is much more helpful than the TS.
You mean you can analyze the output from the TS in the form of a balance and the NS will filter to give "trade or not to trade
So I'll open you eyes wide my friend) The TS also has an output in the form of an error, and you exactly the same way you can train another TS to control the first TS and it will be much more powerful joint
By the way, to trade / not to trade is a "state of the robot" here it is already taken into account
There is a desire to do something like training with an "interactive trainer" (I coined the term myself)
The idea is not to give ready-made labels to AMO, but to teach it by itself from a chart by pointing the mouse "here we should buy with a warrant, and here we should close by market" , you can also teach market conditions, draw attributes (trend lines etc.) , an algorithm that has learned everything in the future should reproduce it all itself ...
https://www.youtube.com/watch?v=tZhkw40gDkE