- This topic has 2 replies, 2 voices, and was last updated 5 years, 7 months ago by
.
-
Posts
-
2 questions:
I was wondering for a while on what’s the “enabled” for in InputVariables. Like, if I don’t need that, why would I type that in the first place? Is it something like commenting out in C for an entire InputVariable? But then would you ever disable that in code? Unless….
I’m trying to dynamically generate a fll file, and for now it’s possible to get an useless InputVariable out of the 5 I have. Detecting this, do I simply disable it? How does this then work with all the rules that include that variable in the middle of the sentence?————————————————
Is there such thing as TOO MANY rules for your engine?
I generated 3125 rules. Is this fine to use both for the engine and performance? In my personal testing, results are inconclusive. May have problems in other parts of my code instead of the rules.If you want to know what I’m doing and maybe could help:
I have 5 InputVariables all ranging from lowest, low, medium, high, highest. Matching with 1 OutputVariable ranging from safe, average, dangerous. I’m trying to think of the inputs as a score: 0, 1, 2, 3, 4. And based on the sum of the 5 inputs, I match with the output.
After a while of trying to think of really complicated rules, I eventually coded a generator to generate 3125 rules for all cases.update: I think things are working now. Had a typo when I originally was struggling when I wrote the questions.
Cool.
Disabled output variables do not defuzzify.
Disabled Input variables do not activate antecedents in the rules (see https://github.com/fuzzylite/fuzzylite/blob/release/fuzzylite/src/rule/Antecedent.cpp#L72).There are no limits on numbers of rules. Personally, I find more than 100 rules to be really hard to trace the correctness of the model. Perhaps you may want to split the model into different controllers and cascade them.
Cheers.
- You must be logged in to reply to this topic.