Machine learning in trading: theory, models, practice and algo-trading - page 1784
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
That feeling when you want to build a model on macroeconomic data (number of U.S. jobs), but stumble upon a coronavirus
Yes, yes our models are retrained...
Here is a link to download the files, even the compressed file does not fit on the forum
https://dropmefiles.com.ua/56CDZB
Train the model on my signs, I wonder what akurasi will be
But if you increase the sample to a year (2019), you can see at least a trend for 2018
From the merging of the two samples we get Accuracy=68.06261099940409
And, from the graph, it appears that the trend of the new trees went sideways.
Decided to train separately on the two samples, since I had to prune them, as you said, the result:
Ваша Accuracy=65.46896295378517
Mine Accuracy=68.03370090447281
That feeling when you want to build a model on macroeconomic data (number of U.S. jobs), but come across a coronavirus
)))
From the merging of the two samples we got Accuracy=68.06261099940409
And, from the graph it turns out that the trend of new trees went sideways.
So your and my signs are together?
Ваша Accuracy=65.46896295378517
Mine Accuracy=68.03370090447281.
You have very nice signs! 3% difference, that's very significant... You probably have your signs tuned for this particular ZZ?
I have also asked how do you trade ketbust in mt4 ?
If i wanted to trade in MT4 with R then i would need to add the signal to a .txt file and let my EA read the file and open/close positions.
)))
So your signs and mine are together?
Yes.
I can give a sample with mine - maybe your method of training will show different results.
That's a great sign you've got! There's a 3% difference, that's very significant... You probably have your signs tuned for this particular zone?
In fact, 3% is within the margin of error, if you make hundreds of thousands of models, as I did before, the variation can be up to 10%.
Yes, I have many target related ZZs. However, I will say that different instruments have different optimal ZZ ranges and if it is selected the result will be better. In fact, ZZ is a return of the price to the past without changing of the trend.
1) And another question, how do you trade ketbust in mt4?
If i want to trade in mt4 using R, for example i may write a signal to tbt file and let my EA from mt4 read the file and open/close a position
1. I trade in MT5. I have a model interpretation class, but it's not mine, so I can't distribute.
2. Yes, of course you can work through a file, that's how many trade copiers work...
Yes.
1) I can give a sample of mine - maybe your method of training will show different results.
2) Actually 3% is within the margin of error, if you make hundreds of thousands of models, as I used to do, the variation can be up to 10%.
3). Yes, of course you can work through a file, that's how many deal copiers work...
1) No, you shouldn't, at least not yet, the main thing is the signs.
2) By experience, as it were - the closer the error is to the limit of model possibilities, the smaller the range of error, if the model gave 55%, the 5% is an error, but if the limit of the model is 68%, the spread is less than 1%
I'm just a complete novice in MT4/5, i have a few ideas to algorithmize, if you would like to participate, i'd be glad to.
1) No, it is not necessary, at least for now, the main thing because the signs.
2) From experience, like - the closer the error is to the limit of model possibilities, the smaller the range of error, if the model gave 55% then the 5% is an error, but if the limit of the model is 68% then the spread is less than 1%
3) What are the names of these kinds of programs / Expert Advisors? How to google them? I'm just a complete novice in MT4/5, I have a few ideas that may be worth algorithmic, if you want to participate, I would be glad to
1. ok.
2. I might have a better one there up to 60% of the time.
3. Yes so look for "Deal Copyer". , that's an option. Depends on the ideas.
Guys, check it out!!!! Don't take it as an advertisement, but I've been to an open house where they told me about the training and you know, in general there were positive results.
First of all, you can learn Python, and secondly, they teach you how to use famous libraries. The presenter, however, lied a little on one of my questions, but it's a mundane thing, but so, who wants to tighten the knowledge in this area please. They also have secured employment.
https://neural-university.ru/main.
But I'm not advertising, just advising only, may the moderators not ban me......
3. Yes so look for "Deal Copyer"
thanks
Depends on the ideas.
Here's one for example.
A normal trend algorithm, no parameters, no tweaks , there is only one sweep that smoothes one signal line
If you do not like it, you will get a very good result, but I will tell you once again, the robot has no parameters, no optimizations, no MO, only adaptivity and spectral analysis. You cannot do that with ordinary indicators.
If you want to realize such a robot you should write in MT4 the principal components method and the calculation of the Pearson correlation coefficient.
If you do not know what to do with R, you cannot do it with mt4.
thanks
Here is one of them, for example.
The usual trend algorithm, no parameters, no tweaks , there is only one sweep that smoothes one signal line
I think this robot has a very good effect, but I want to reiterate that it has no parameters and no optimizations, only adaptivity and spectral analysis. You can't do that with ordinary indicators.
If you want to realize such a robot you should write in MT4 the principal components method and the calculation of the Pearson correlation coefficient.
If you do not know how to do it in R then it will be a noobrainer.
That feeling when you want to build a model on macroeconomic data (number of U.S. jobs) but run into a coronavirus
One external factor can't be significant on a large scale. You have to take a group of nearby ones, at least a couple more. That would be more accurate. Coronavirus cannot be taken into account, although you can / should compare it with crises. At least highlight which factors are still significant.