Machine learning in trading: theory, models, practice and algo-trading - page 2806
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
Ah, well, standard indicators and their derivatives mostly?
I initially used my trading experience, my vision of price interaction with different indicators, price levels and other patterns - something confirmed and something not. That said, I'm more complaining about the model, that it throws out reliable but rare events.....
I used to be sceptical about oscillators, but recent experiments have shown that they have stable signals, for example MACD.
to train separately on rising and falling markets? Then the signals will be averaged anyway.
in general everything is logical, counter-trend trades are usually ineffective in crack markets, with some scalper exceptions.
I just showed you that you can get a better understanding of how the model works.
Why average, there will be 3 models - one determines which of the two models to use.
Did you get the same profitability for buying and selling?
Yes, you've shown correctly that the effectiveness of the model depends on the data, of course.
Why average, there will be 3 models - one determines which of the two models to use.
Have you got the same profitability for buying and selling?
Yes, it was correctly shown that the effectiveness of the model depends on the data, of course.
I haven't looked at the statistics on the trades yet.
3 models is interesting, by the way, I only have 2 so far... it makes sense.
There have been several posts above about not necessarily removing correlated predictors.
I cannot accept the justification that the model algorithm is robust to correlated predictors.
Yes, the algorithm is robust if there are a few pieces of correlated predictors in a set of a hundred predictors.
But what if all predictors are correlated? What if most are correlated? Where is the boundary?
Removing correlated predictors is to reveal the quality of the set of predictors, and the characteristics of a particular model's algorithm with respect to correlation are completely unimportant. Before modelling, one should just know the stricti model on one predictor or on a hundred. It is necessary to know the number of predictors on which the model is built.
Started your script, and it has some moments that prevent its further operation:
1. A comma instead of a full stop
2. The last columns are lost.
Can you fix it?
Ran your script and it has some issues preventing further use:
1. A comma instead of a full stop
2. the last columns are lost.
Can you fix it?
Can you be more specific?
Can you be more specific?
Here's the right-hand piece of the sample in table form
Below is the table of the script results in the same location
As you can see, there are no information columns, including target columns.
And these columns are at the very beginning of the file, as it turned out
About comma instead of dot as a number separator, either I made a mistake or corrected it.
Your script has been running for more than a day and has not yet created a single file based on the results of the screening. I don't know, maybe it's time to switch it off?
Can you be more specific?
I swapped it around and it seemed fine.
I swapped it around and it seemed fine.
.