Jerry Mendel’s book can be safely considered as the Type-2 Fuzzy Logic bible. It is not an easy or inexpensive read, but definitely the best way to get to know the world of Type-2 Sets.
In this post a book example, where the centroid of a number of Interval Type-2 sets was calculated is replicated using the Type-2 Library.
Type reduction in these cases is carried out by using the function it2_kernikmendel_reduce(interval_set), and a crisp set, the centroid, is returned back.
'''
This module executes the karnik-mendel type reduction
algorithm and compares the results with those obtained
in mendel's book
References:
-----------
Mendel, Jerry M. Uncertain rule-based fuzzy systems.
Springer, Cham, 2017. pages 261-262
'''
import numpy as np
from numpy.random import normal
import math
import matplotlib.pyplot as plt
from type2fuzzy import IntervalType2FuzzySet
from type2fuzzy import it2_kernikmendel_reduce
def gaussian(x, mean, sigma):
g = np.exp(-0.5*(((x - mean)/sigma)**2))
return g
def generate_sets_uncertain_mean(m1_m2_list):
it2fs_list = []
x= np.linspace(0,10,101)
for m1_m2 in m1_m2_list:
m1 = m1_m2[0]
m2 = m1_m2[1]
g1 = gaussian(x, m1 ,1)
g2 = gaussian(x, m2 ,1)
hmf = np.maximum(g1,g2)
one_indexes = np.where(hmf==1)[0]
if len(one_indexes) > 1:
hmf[one_indexes[0]: one_indexes[1]] = 1
lmf = np.minimum(g1,g2)
it2fs_list.append(
IntervalType2FuzzySet.from_hmf_lmf(x, hmf, lmf))
return it2fs_list
def generate_sets_uncertain_variance(m1_m2_list):
it2fs_list = []
x= np.linspace(0,10,101)
for s1_s2 in m1_m2_list:
s1 = s1_s2[0]
s2 = s1_s2[1]
g1 = gaussian(x, 5 ,s1)
g2 = gaussian(x, 5 ,s2)
hmf = np.maximum(g1,g2)
lmf = np.minimum(g1,g2)
it2fs_list.append(
IntervalType2FuzzySet.from_hmf_lmf(x, hmf, lmf))
return it2fs_list
def it2fs_centroid(it2fs_list):
for it2fs in it2fs_list:
centroid = it2_kernikmendel_reduce(
it2fs, information='none', precision=4)
print(f'Centroid: {centroid}')
# test 1 - table 9-1
# results obtained:
# [5.00000]
# [4.87498, 5.12502]
# [4.74952, 5.25048]
# [4.62265, 5.37735]
# [4.49285, 5.50715]
# [4.21675, 5.78325]
# [3.90697, 6.09303]
# [3.55194, 6.44806]
# [3.15053, 6.84947]
m1_m2_list = [(5,5), (4.875,5.125), (4.75, 5.25),
(4.625, 5.375), (4.5, 5.5), (4.25, 5.75), (4,6),
(3.75, 6.25), (3.5 ,6.5)]
it2fs_list_m = generate_sets_uncertain_mean(m1_m2_list)
print('Uncertain Mean Results')
it2fs_centroid(it2fs_list_m)
print('\n')
# test 2 - table 9-2
# results obtained:
# [5.00000]
# [4.80054, 5.19946]
# [4.60079, 5.39921]
# [4.39849, 5.60151]
# [4.18488, 5.81512]
# [3.93441, 6.06559]
# [3.59388, 6.40612]
s1_s2_list = [(1,1), (0.875,1.125), (0.75, 1.25),
(0.625,1.375), (0.5, 1.5), (0.375,1.625), (0.25,1.75)]
it2fs_list_v = generate_sets_uncertain_variance(s1_s2_list)
print('Uncertain Variance Results')
it2fs_centroid(it2fs_list_v)