Skip to content

Instantly share code, notes, and snippets.

@huberflores
Last active April 2, 2022 22:43
Show Gist options
  • Star 2 You must be signed in to star a gist
  • Fork 1 You must be signed in to fork a gist
  • Save huberflores/f5df82be92b05f7e5c7e to your computer and use it in GitHub Desktop.
Save huberflores/f5df82be92b05f7e5c7e to your computer and use it in GitHub Desktop.
Fuzzy logic in R
#
# author Huber Flores
#
# Example of fuzzy logic in R
# The main idea is to have fuzzy sets for each of the dimensions
# each attribute in the dimension can be defined as a fuzzy set with their respective linguistic variables, terms and membership functions
library(sets)
U1 <- seq(from = 0, to = 1, by = 0.0001)
#DIMENSIONS
variables <-
set(
code =
fuzzy_partition(varnames=
c(ELow = 0.2 , ENormal = 0.5, EHigh = 0.8),
FUN= fuzzy_cone, radius = 0.2, universe=U1),
bandwidth =
fuzzy_partition(varnames=
c(SLow = 0.2, SNormal=0.5, SHigh=0.8),
FUN = fuzzy_cone, radius = 0.2, universe=U1),
acceleration =
fuzzy_partition(varnames=
c(ALow = 0.2, ANormal=0.5, AHigh=0.8),
FUN = fuzzy_cone, radius = 0.2, universe=U1),
processing =
fuzzy_partition(varnames=
c(Local = 0.3, Remote = 0.7),
FUN = fuzzy_cone, radius = 0.3, universe=U1)
)
rules <-
set(
fuzzy_rule(code %is% EHigh && bandwidth %is% SHigh && acceleration %is% AHigh, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SHigh && acceleration %is% ANormal, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SHigh && acceleration %is% ALow, processing %is% Local),
fuzzy_rule(code %is% EHigh && bandwidth %is% SNormal && acceleration %is% AHigh, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SNormal && acceleration %is% ANormal, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SNormal && acceleration %is% ALow, processing %is% Local),
fuzzy_rule(code %is% EHigh && bandwidth %is% SLow && acceleration %is% AHigh, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SLow && acceleration %is% ANormal, processing %is% Local),
fuzzy_rule(code %is% EHigh && bandwidth %is% SLow && acceleration %is% ALow, processing %is% Local),
fuzzy_rule(code %is% EHigh && bandwidth %is% SHigh, processing %is% Remote),
fuzzy_rule(code %is% EHigh && bandwidth %is% SLow, processing %is% Local),
fuzzy_rule(code %is% EHigh && bandwidth %is% SNormal, processing %is% Local),
fuzzy_rule(code %is% ENormal && bandwidth %is% SLow, processing %is% Local),
fuzzy_rule(code %is% ENormal && bandwidth %is% SHigh, processing %is% Remote),
fuzzy_rule(code %is% ENormal && bandwidth %is% SNormal, processing %is% Local),
fuzzy_rule(code %is% ELow && bandwidth %is% SLow, processing %is% Local),
fuzzy_rule(code %is% ELow && bandwidth %is% SNormal, processing %is% Local),
fuzzy_rule(code %is% ELow && bandwidth %is% SHigh, processing %is% Local)
)
context <- fuzzy_system(variables, rules)
print(context)
#plot(context)
#fi <- fuzzy_inference(context, list(code=0.8123, bandwidth=0.912))
#0.7
#fi <- fuzzy_inference(context, list(code=0.5212, bandwidth=0.2121))
#0.3
#fi <- fuzzy_inference(context, list(code=0.1532, bandwidth=0.9321))
#0.3
#fi <- fuzzy_inference(context, list(code=0.2432, bandwidth=0.9323))
#0.3
#fi <- fuzzy_inference(context, list(code=0.4723, bandwidth=0.9542))
#0.7
#fi <- fuzzy_inference(context, list(code=0.2932, bandwidth=0.9484))
#0.3
#fi <- fuzzy_inference(context, list(code=0.2912, bandwidth=0.4324))
#0.3
#fi <- fuzzy_inference(context, list(code=0.2134, bandwidth=0.2194))
#0.3
#fi <- fuzzy_inference(context, list(code=0.4323, bandwidth=0.5345))
#0.3
#fi <- fuzzy_inference(context, list(code=0.8123, bandwidth=0.912, acceleration = NA))
#There is a rule code is High AND bandwidth is High -> Remote.
#as well as a rule, code is High AND bandwidth is High and acceleration is High -> Remote.
#0.7
#fi <- fuzzy_inference(context, list(code=0.8123, bandwidth=0.912, acceleration = 0.93))
#0.7
#fi <- fuzzy_inference(context, list(code=0.8123, bandwidth=0.912, acceleration = 0.13))
#0.5
#fi <- fuzzy_inference(context, list(code=0.8123, bandwidth=0.2354, acceleration = 0.9243))
#0.4517523
#fi <- fuzzy_inference(context, list(code=0.4723, bandwidth=0.9542, acceleration=NA))
#This rule exists for the two variables provided
#0.7
#fi <- fuzzy_inference(context, list(code=0.8857, bandwidth=0.1194, acceleration=0.8921))
#0.4961897
fi <- fuzzy_inference(context, list(code=0.8857, bandwidth=0.1194, acceleration=0.1184))
#0.3
#dev.new()
#plot(fi)
gset_defuzzify(fi, "centroid")
U1 <- NULL
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment