Machine learning in trading: theory, models, practice and algo-trading - page 762
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
Here is the balance in two weeks, this is the third. The same model is working. Last week there was a change of futures, so there were few trades, as soon as I changed the futures, it started working. Yesterday I caught two craps and immediately stopped trading. Today I restarted it in the evening, although I have missed one good signal. But on the whole, everything is good so far. I wish you that.
eb.... Is this real? $1000 in 2 weeks.... it's not like i have 20 - 50 bucks in my account max.... I'm just a guy from Togliatti, I could only dream of such money... Imho very good. By the way Michael, about the indicator it was not the babine))))) in the turntable on XP x32 everything worked. So, do not worry, I don`t need any dlls and induks, I use what I have...
I think I'll do something by the end of the week...
No sunshine
Gloomy spring
Drain in the darkness...
Here is the balance in two weeks, this is the third. The same model is working. Last week there was a change of futures, so there were few trades, as soon as I changed the futures, it started working. Yesterday I caught two craps and immediately stopped trading. Today I restarted it in the evening, although I have missed one good signal. But on the whole, everything is good so far. What I wish you...
I'll make a new roll in the next few days and start monitoring, time to show you how it's done (or not).
eb.... Is this real? $1,000 in 2 weeks.... it's not like i have 20 - 50 bucks in my account max.... Well I'm just a guy from Togliatti I can only dream about that kind of money... Imho very good.
А? What? Is the grail around here somewhere? Who is the last in line?
Here is the balance in two weeks, this is the third. The same model works. Last week there was a change of futures, so there were few trades, as soon as I changed the futures, it started working. Yesterday I caught two bad trades and had to urgently stop trading. Today I restarted it in the evening, although I have missed one good signal. But on the whole, everything is good so far. I wish you to do the same.
If such a big increase from 1 deal - i guess you trade with a big part of your deposit? For example, do you trade 50% of the deposit and with a good leverage?
The percentage of profit is not the right indicator, the main thing is profitability and equity type. In the beginning, indeed, the load on the deposit is very large..... But the fact itself, the model is gaining so far. I think I will hold it till the end of the week and then I will re-train it....
Michael, can you give me some advice as someone with experience. I have about 1000 input data. But with 1000 predictors even NeuroSolution hangs. What do you think if I divide 1000 inputs into 5 groups of 200 inputs. Each group is a separate network in the end 5 networks. Then I feed the outputs from the networks to the 6th network, the output of which already gets the desired value. The question is how do you think it will work or should I create one network but with all 1000 inputs at once?
Michael, can you give me some advice as someone with experience. I have about 1000 input data. But with 1000 predictors even NeuroSolution hangs. What do you think if I divide 1000 inputs into 5 groups of 200 inputs. Each group is a separate network in the end 5 networks. Then I feed the outputs from the networks to the 6th network, the output of which already gets the desired value. The question is how do you think it will work and whether to create a single network, but with all 1000 inputs at once.
If the input data is a lot of 1000 for one answer and a little 1000 such examples, most likely the significance of the variables in the example is very weak and noisy. the number of examples I personally use from 10 000, and with such a number well enough to handle the algorithm random forrest. and answering your question about the results of 5 grids then send to the 6th grid yes, so you can train. But even here we can take a slightly different approach, just put the importance of these 5 grids on the ballot. i.e. sort of a summary forecast to use.
It is simpler to eliminate out of 1000 inputs and leave only those which affect the output. I'm sure there will be not more than 50 and so on. After that everything is on the same network. Selection is done using vtreat.R
Is it possible to do as I described above? I know that there are committees from networks, but I've never worked with them. And I assume that training should be carried out in parallel... Or it is still possible to separately? I am interested in your opinion.