-
Notifications
You must be signed in to change notification settings - Fork 0
/
classificador.py
219 lines (166 loc) · 6.41 KB
/
classificador.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
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
from scipy.spatial import distance
from matplotlib import pyplot as plt
import statistics
import copy
class classificador:
def __init__(self, pathZ1, pathZ2, pathZ3):
self.z1 = self.getSamples(pathZ1)
self.z2 = self.getSamples(pathZ2)
self.z3 = self.getSamples(pathZ3)
def generateBoxplot(self, method1List, method2List, path):
data = []
data.append(method1List)
data.append(method2List)
box = plt.boxplot(data, patch_artist=True)
colors = ['lightgreen', 'tan']
for patch,color in zip(box['boxes'],colors):
patch.set_facecolor(color)
plt.xticks([1, 2], ['Método 1', 'Método 2'])
plt.ylabel('Erros por execução')
plt.savefig(path)
plt.clf()
def generateGraphs(self, yList1, yList2, xList, path):
xList = [i+1 for i in xList]
plt.figure(figsize=(12, 6))
plt.plot(xList, yList1, 'b--', marker='o', label='Validação')
plt.plot(xList, yList2, 'r--', marker='o', label='Teste')
plt.legend(loc='upper right')
plt.xticks(xList)
aux1 = max(yList1)
if aux1 < max(yList2):
aux1 = max(yList2)
aux2 = min(yList1)
if aux2 > min(yList2):
aux2 = min(yList2)
plt.ylim(0, aux1+(aux1*0.2))
#plt.xlim(0, max(xList))
plt.ylabel('Probabilidade média dos erros por execução (%)')
plt.xlabel('Número de iterações')
plt.savefig(path)
plt.clf()
def getSamples(self, pathTxt):
with open(pathTxt) as f:
samples = f.readlines()
for i in range(0, len(samples)):
samples[i] = samples[i].replace('[', '').replace(']', '').replace('\n', '').replace('\'', '').replace(' ', '').split(',')
for i in range(0, len(samples)):
for j in range(0, len(samples[i])):
try:
samples[i][j] = float(samples[i][j])
except ValueError:
samples[i][j] = samples[i][j]
return samples
def calculateDist(self, idClass, train=True):
base_busca = copy.deepcopy(self.z1)
if train == True:
base_ref = copy.deepcopy(self.z2)
elif train == False:
base_ref = copy.deepcopy(self.z3)
for i in range(0,len(base_ref)):
#print(base_ref[i])
base_ref[i].pop(idClass)
for i in range(0,len(base_busca)):
#print(base_busca[i])
base_busca[i].pop(idClass)
distList = []
for rs in range(0,len(base_ref)):
distList.append([])
for bs in range(0,len(base_busca)):
distList[rs].append([])
distList[rs][bs].append(distance.euclidean(base_ref[rs], base_busca[bs]))
distList[rs][bs].append(self.z1[bs])
return distList
def calculateError(self, idClass, K, train=True):
distList = self.calculateDist(idClass, train=train)
nErrors = 0
sErrors = []
for i in range(0,len(distList)):
distList[i] = sorted(distList[i])
flag=1
kn = K
while(flag==1):
try:
flag=0
kClasses = []
for k in range(kn):
kClasses.append(distList[i][k][1][idClass])
attrClass = statistics.mode(kClasses)
except statistics.StatisticsError:
kn=kn-1
flag=1
if train == True:
if self.z2[i][idClass] != attrClass:
nErrors=nErrors+1
sErrors.append(self.z2[i])
elif train == False:
if self.z3[i][idClass] != attrClass:
nErrors=nErrors+1
sErrors.append(self.z3[i])
return nErrors,sErrors
def findBestK(self, idClass, kList, train=True):
minNErrors = None
xList = []
yList = []
for k in kList:
nErrors,sErrors = self.calculateError(idClass, k, train=train)
xList.append(k)
yList.append(nErrors)
if minNErrors == None or nErrors < minNErrors:
bestK = k
minNErrors = nErrors
minSErrors = sErrors
return minNErrors,minSErrors,bestK
if __name__=='__main__':
import classificador as cl
kList = []
kList.append(1)
kList.append(3)
kList.append(5)
kList.append(9)
kList.append(11)
kList.append(13)
kList.append(15)
kList.append(17)
kList.append(19)
kList.append(21)
kList.append(23)
pathZ1 = 'data/ecoli.data/z1.txt'
pathZ2 = 'data/ecoli.data/z2.txt'
pathZ3 = 'data/ecoli.data/z3.txt'
obj = cl.classificador(pathZ1, pathZ2, pathZ3)
NErrors,SErrors,bestK = obj.findBestK(idClass=7, kList=kList, train=True)
print('bestK: %i' % bestK)
print('NErrors: %i' % NErrors)
print('SErrors: %s\n' % str(SErrors))
pathZ1 = 'data/iris.data/z1.txt'
pathZ2 = 'data/iris.data/z2.txt'
pathZ3 = 'data/iris.data/z3.txt'
obj = cl.classificador(pathZ1, pathZ2, pathZ3)
NErrors,SErrors,bestK = obj.findBestK(idClass=4, kList=kList, train=True)
print('bestK: %i' % bestK)
print('NErrors: %i' % NErrors)
print('SErrors: %s\n' % str(SErrors))
pathZ1 = 'data/seeds.data/z1.txt'
pathZ2 = 'data/seeds.data/z2.txt'
pathZ3 = 'data/seeds.data/z3.txt'
obj = cl.classificador(pathZ1, pathZ2, pathZ3)
NErrors,SErrors,bestK = obj.findBestK(idClass=7, kList=kList, train=True)
print('bestK: %i' % bestK)
print('NErrors: %i' % NErrors)
print('SErrors: %s\n' % str(SErrors))
pathZ1 = 'data/wine.data/z1.txt'
pathZ2 = 'data/wine.data/z2.txt'
pathZ3 = 'data/wine.data/z3.txt'
obj = cl.classificador(pathZ1, pathZ2, pathZ3)
NErrors,SErrors,bestK = obj.findBestK(idClass=0, kList=kList, train=True)
print('bestK: %i' % bestK)
print('NErrors: %i' % NErrors)
print('SErrors: %s\n' % str(SErrors))
pathZ1 = 'data/yeasts.data/z1.txt'
pathZ2 = 'data/yeasts.data/z2.txt'
pathZ3 = 'data/yeasts.data/z3.txt'
obj = cl.classificador(pathZ1, pathZ2, pathZ3)
NErrors,SErrors,bestK = obj.findBestK(idClass=8, kList=kList, train=True)
print('bestK: %i' % bestK)
print('NErrors: %i' % NErrors)
print('SErrors: %s\n' % str(SErrors))