-
Notifications
You must be signed in to change notification settings - Fork 1
/
TSK.py
175 lines (162 loc) · 8.02 KB
/
TSK.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
import pandas as pd
import numpy as np
import math as math
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
import statistics
import csv
#-----------------------------------------------------------------------------------------------------------------------------#
#------------------------------VIRTUAL FLOW SENSOR DESIGN USING TAKAGI-SUGENO-KANG METHOD----------------------------------------#
#-----------------------------------------------------------------------------------------------------------------------------#
df = pd.read_csv('data_zawor_2001-1.csv', sep=',')
data=np.array( [ df["X[%]"][0:1800],df["P1[kPa]"][0:1800]-df["P2[[kPa]"][0:1800], df["F[m^3/h]"][0:1800]]) # learn dataset
val_data=np.array( (df["X[%]"][1801:], (df["P1[kPa]"][1801:]-df["P2[[kPa]"][1801:]), df["F[m^3/h]"][1801:])) # validation dataset
template=["S0","CE","B0"] # names of triangular functions of membership
regions=3 # number of regions
N=1
peaks=np.zeros((3,regions)) # points in datasets of maximum value of membership functions
predictions= []
def allpeaks(data): #definition of all peaks of membership functions for a given input/output data
for i in range (len(data)):
max_v=max(data[i])
min_v=min(data[i])
for j in range (regions):
peaks[i][j]=min_v+ j * (max_v-min_v)/(2*N)
def d_degree (data,number, implication): # functions of membership, blurring
one_series=[]
for record in data:
membership = np.zeros((regions, 1))
for i in range (regions-1):
if (record >= peaks[number, i] and record <= peaks[number, i + 1]):
membership[i][0] = (peaks[number, i + 1] - record) / (peaks[number, i + 1] - peaks[number, i])
membership[i + 1][0] = (record - peaks[number, i]) / (peaks[number, i + 1] - peaks[number, i])
break
index = np.where(membership != 0)[0]
if ( implication == True and len(index) == 2): # Picking maximum value of function of membership
if (membership[index[0]][0] > membership[index[1]][0]):
membership[index[0]][0] = 0;
else:
membership[index[1]][0] = 0
one_series.append(membership)
return one_series
def all_degres(data,implication): # for each input dataset genereting fuzzy values
series=0
all_membership=[]
for i in data:
all_membership.append(d_degree(i,series,implication))
series+=1
return all_membership
def set_rules (input, list_of_mem): # A set of rules, returning sharp exits of specific premises
y_all= []
for i in range(len(data[0])):
y= {}
if (list_of_mem[0][i][0] != 0 and list_of_mem[1][i][0] != 0):
y[0] = max(data[2]) - 0 * input[0][1] - 0.01 * input[1][i] ** 0.5
if (list_of_mem[0][i][0] != 0 and list_of_mem[1][i][1] != 0):
y[1] = max(data[2]) - 0.05 * input[0][1] - 0.1 * input[1][i] ** 0.5
if (list_of_mem[0][i][0] != 0 and list_of_mem[1][i][2] != 0):
y[2] = max(data[2]) - 0 * input[0][1] - 0 * input[1][i] ** 0.5
if (list_of_mem[0][i][1] != 0 and list_of_mem[1][i][0] != 0):
y[3] = max(data[2]) - 0.55 * input[0][1] - 0.5 * input[1][i] ** 0.5
if (list_of_mem[0][i][1] != 0 and list_of_mem[1][i][1] != 0):
y[4] = max(data[2]) - 0.5 * input[0][1] - 0.5 * input[1][i] ** 0.5
if (list_of_mem[0][i][1] != 0 and list_of_mem[1][i][2] != 0):
y[5] = max(data[2]) - 0.80 * input[0][1] - 0.8 * input[1][i] ** 0.5
if (list_of_mem[0][i][2] != 0 and list_of_mem[1][i][0] != 0):
y[6] = max(data[2]) - 1.15 * input[0][1] - 1.5 * input[1][i] ** 0.5
if (list_of_mem[0][i][2] != 0 and list_of_mem[1][i][1] != 0):
y[7] = max(data[2]) - 1 * input[0][1] - 0.71 * input[1][i] ** 0.5
if (list_of_mem[0][i][2] != 0 and list_of_mem[1][i][2] != 0):
y[8] = max(data[2]) - 1.3 * input[0][1] - 1.2 * input[1][i] ** 0.5
y_all.append(y)
print("--------")
return y_all
def switchfun(i,memX1, memX2): # A function that returns the result of Mandami implication
switcher={
0: min(memX1[0],memX2[0]),
1: min(memX1[0],memX2[1]),
2: min(memX1[0],memX2[2]),
3: min(memX1[1],memX2[0]),
4: min(memX1[1],memX2[1]),
5: min(memX1[1],memX2[2]),
6: min(memX1[2],memX2[0]),
7: min(memX1[2],memX2[1]),
8: min(memX1[2],memX2[2]),
}
return switcher.get(i)
def deffuzying (output, memberships): # defuzzification
for record in range (len(data[0])):
counter_sum=0
denom_sum = 0
for i in output[record]:
min_val= switchfun(i,memberships[0][record],memberships[1][record])
counter_sum+=min_val*output[record][i]
denom_sum+=min_val
predictions.append(float(counter_sum/denom_sum))
def errocount (real): # Wyznaczanie statystyki błędów
l = []
l = (real[2] - predictions) / max(data[2]) * 100
avarage_error = sum(abs(l)) / len(l)
max_error = max(l)
min_error = min(l)
print('Minimal error: ' + str(min_error))
print('Maximum error: ' + str(max_error))
print('Avarage error: ' + str(avarage_error))
print("Standard Deviation of sample: " + str(statistics.stdev(predictions)))
print('Statistic harmonic mean: ' + str(statistics.harmonic_mean(predictions)))
return l
# -----------------------------------------------------------------------------------------------------------------------------#
# ---------------------------------------------------------PLOTS---------------------------------------------------------------#
# -----------------------------------------------------------------------------------------------------------------------------#
def controlPane_draw (data1,data2,data3, tittle, xname, yname, zname):
fig = plt.figure()
ax = fig.gca(projection='3d')
ax.plot_trisurf(data1, data2, data3, linewidth=0.2, antialiased=True, cmap='plasma')
ax.set_xlabel(xname)
ax.set_ylabel(yname)
ax.set_zlabel(zname)
fig.suptitle(tittle)
plt.show()
def histogram_draw (data , tittle, xname, yname):
x=data
bins_num= 80
plt.hist(x, bins= bins_num)
plt.title(tittle)
plt.ylabel(yname)
plt.xlabel(xname)
plt.show()
def one_plot (data1, data2 , tittle, yname, xname, plotname1, plotname2):
x=np.linspace(0,1800, num = 1800)
plt.figure(figsize=(12, 5))
plt.plot(x, data1, label=plotname1)
if data2 != 0:
plt.plot(x, data2, label=plotname2)
plt.xticks(np.arange(0, 1900, 100))
plt.title(tittle)
plt.ylabel(yname)
plt.xlabel(xname)
plt.legend(loc="lower right")
plt.show()
def allPlotsdraw (X1,X2,Y_real,Y_predictions,error): # funkcja rysująca wszystkie wykresy
controlPane_draw(X1,X2,Y_real, "Plane of real control", "X[%]", "dP[[kPa]", "F[m^3/h]")
controlPane_draw(X1,X2,Y_predictions, "Plane of predicted control", "X[%]", "dP[[kPa]", "F[m^3/h]")
histogram_draw(Y_real, 'Histogram of Real values', 'Predictions', 'Frequency')
histogram_draw(Y_predictions, 'Histogram of predictions', 'Predictions', 'Frequency')
one_plot(Y_real, Y_predictions, 'Takagi-Sugeno-Kang results', 'F[m^3/h]', 'time[s]', 'Real', " Predictions")
one_plot(error, 0, 'Relative error', '[%]', 'time[s]', 'error', '')
def runProject(): # Funkcja wykonująca wszystkie potrzebne do projektu operacje
allpeaks(data)
#test on learn dataset
list_of_mem = all_degres(data, False)
y = set_rules(data, list_of_mem)
deffuzying(y, list_of_mem)
error=errocount(data)
allPlotsdraw(data[0],data[1],data[2],predictions,error)
#Test on validation dataset
predictions.clear()
list_of_mem = all_degres(val_data, False)
y = set_rules(val_data, list_of_mem)
deffuzying(y, list_of_mem)
error = errocount(val_data)
allPlotsdraw(val_data[0], val_data[1], val_data[2], predictions, error)
runProject()