<- Back to Tasks

improve fuzzy logic settings

if i want to try some fuzzy logic strategies, where i want 1-5 conditions, i need to set conditions to 5-5 and fuzzy logic from 21% to 95%, so it will compute conditions as 1,2,3,4,5



but why the hell need the genetics go from 21% to 95% and try every possible combination between 21-95 - we are slowing down the process and trying more and more combinations which will lead to the same number of conditions


so if i set i want to generate fuzzy logic strategie, i could set only set min/max conditions for exampl 1-5 and i can choose what ever combination i want - 2,3,4, ...

Attachments
fuzzy.jpg
fuzzy.jpg
(171.67 KiB)
  • Votes +2
  • Project StrategyQuant X
  • Type Feature
  • Status Refused
  • Priority Normal

History

h
#1

hankeys

06.02.2019 08:58

Task created

m
#2

mabi

06.02.2019 09:04
Voted for this task.
KL
#3

kainc301

22.02.2019 12:48

Type changed from Bug to Feature

t
#4

todtv

23.02.2019 12:19
Voted for this task.
MF
#5

Mark Fric

29.04.2019 10:42

Status changed from New to Waiting for information

I don't understand what you want to achieve.


These are two different things;

- Number of conditions generated 

- How many of these conditions have to be true for fuzzy signal to be true

h
#6

hankeys

29.04.2019 12:34
je prostě blbost nastavovat min a max kondici v procentech od do, která je pak v SQX používána s krokem 1

stačilo by nastavit hodnotu max kondice a když nastavím třeba 5, tak se nabídnou checkboxy v rozsahu 1-5 a já už si vyberu, že chci např. vybírat kombinaci 2,3,4,5 parametrů

abych tohoto dosáhl, tak musím nastavit 5-5, min 21% max 81% a do genetiky tak pak zkouší v rozsahu 21-81, tzn. 60 kroků - mrháme časem a parametry strategie neměníme, protože je jedno, jestli je min kondice 21% nebo 22%, budou vždycky použity 2 parametry


proč? aby se nemrhalo genetikou na změnu parametrů kondice, která skáče v procentech s krokem 1

MF
#7

Mark Fric

29.04.2019 12:58

Status changed from Waiting for information to Refused

misunderstood, it has no efect in number of steps in genetics, and it is not less effective with percentages. 


It doesn't matter if the number of true conditions is set as numbers or as percentage, they are randomly generated once. 

It also does not matter that 21% is the same number of conditions as 22, 23, 24, 25%, that number is generated only once.


Percentage values ensure that it will be always valid no matter how many conditions you have, without it we wouldnt be able to combine strategies with different number of conditions.

h
#8

hankeys

29.04.2019 13:12
OK
Votes: +2

Drop files to upload

or

choose files

Max size: 5MB

Not allowed: exe, msi, application, reg, php, js, htaccess, htpasswd, gitignore

...
Wait please