Discussion of article "Fuzzy Logic in trading strategies" - page 2
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
To divide the "trading space" into zones of fuzzy logic. is not in itself a manifestation of "crisp" logic:)
for this purpose it is possible to shift belonging functions in the optimiser, + it is possible to build complex combined functions, which are difficult to comprehend in general :) but in the optimiser to pick up interesting sets
Thank you!
Very interesting article.
I have been wanting to understand the topic for a long time.
Today my robot makes trading decisions on the basis of consensus modules each of which works with its own pattern. Since there are quite a lot of them, I select the necessary set by the method of disconnection. I think it would be interesting to apply fuzzy logic to the set.
Thank you!
Very interesting article.
I have been wanting to understand the topic for a long time.
Today my robot makes trading decisions on the basis of consensus modules each of which works with its own pattern. Since there are quite a lot of them, I select the necessary set by the method of disconnection. I think it would be interesting to apply fuzzy logic to the set.
Please, yes, this is exactly the right case to try.
Very interesting topic, thanks to the author!
Didn't like this line with all the consequences:
#include <MT4Orders.mqh>It's not clear why pull in homemade libraries from MT4?
Then the coding style... Where are the comments in the code?
Very interesting topic, thanks to the author!
Didn't like this line with all the consequences:
It's not clear why pull in homemade libraries from MT4?
This is a library for MT5 to work with orders in MT5 just as in MT4, very simplifies the transfer of code to MT4 (if necessary), in my opinion very well done, just always use it myself. It simplifies life in general.
Then the coding style... Where are the comments in the code?
I forgot about comments :) I have commented all the main things in the article.
Maxim, one more thing, if I may have an opinion.
As far as I understand, the idea of the article is to show the advantage of fuzzy logic. The nuance is that this methodology (obtaining a conclusion) is implemented step by step. So, in the code of the Expert Advisor Fuzzy logic for fuzzy algotraders.mq5 I would indicate these stages in the comments. So that then you can use them as a template for any EA.
Yes, hopefully the author will continue this topic. Thanks again.
Thanks to the author for good work, clear presentation with a practical example.
IMHO only, I would like a more convincing justification of the advantages of the proposed method, because the comparison of the optimised balance chart for an EA on fuzzy logic with a non-optimised one for an EA on clear logic is not quite correct.
Although the author wrote that it is very much and difficult to select optimisation criteria for clear rules, but I tried it and it came out exactly two lines in Expert_without_fuzzy.mq5 EA:
ie added the possibility of optimising the upper and lower levels of the indicator.
And here is the result of the test on the same period as in the article from 01.01.2017 and timeframe - M15:
it seems that now the result is not so different from the similar one with fuzzy logic....
Thanks to the author for good work, clear presentation with a practical example.
IMHO only, I would like a more convincing justification of the advantages of the proposed method, because the comparison of the optimised balance chart for an EA on fuzzy logic with a non-optimised one for an EA on clear logic is not quite correct.
Although the author wrote that it is very much and difficult to select optimisation criteria for clear rules, but I tried it and it came out exactly two lines in Expert_without_fuzzy.mq5 EA:
ie added the possibility of optimising the upper and lower levels of the indicator.
And here is the result of the test on the same as in the article period from 01.01.2017 and timeframe - M15:
it seems that now the result is not so different from the similar one with fuzzy logic....
Excellent point, you are right, but in fuzzy logic the result also depends on the membership functions. Suppose we don't know which indicator readings to use to buy and which to sell, then we can swap the extreme belonging functions, or we can make them triangular or combined - here it will be more difficult to rewrite your example and we will have to make sure that the clear logical conditions are not suddenly contradictory..., let's say we make a triangular accessory f-ia for the term "sell" with a 0.7 centre, and closer to 1 it will decrease instead of increasing, and for all 3 indicators we will put different f-ia and change the terms "buy" and "sell" places, then the results will change not linearly and many new situations will arise. Whereas in your code it will be necessary to strictly control that the conditions do not contradict each other, otherwise the f-ya will return only the result of the last condition in the code. This is the first thing that came to mind.
But in general, as adepts of fuzzy logic say - I heard in some video that everything can be done without it, but why if in some cases it is more convenient, somehow so :).
Loved the article! Write more.
Who can explain in words what Mamdani and Sugeno algorithms do?