-
Notifications
You must be signed in to change notification settings - Fork 0
/
passfail.py
163 lines (136 loc) · 5.63 KB
/
passfail.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
#%%
# 載入需要的package
import pandas as pd
import numpy as np
import seaborn as sns # 畫圖用
import matplotlib.pyplot as plt # 畫圖用
from imblearn.over_sampling import SMOTE #oversampling
from imblearn.under_sampling import NearMiss #underersampling
from sklearn.model_selection import train_test_split # 分割訓練集與測試集
from sklearn.metrics import classification_report # recall precision...報告
from sklearn.metrics import confusion_matrix # 混淆矩陣
from sklearn.linear_model import LogisticRegression # 邏輯式迴歸模型
from sklearn import svm # svm模型
# 讀檔
data = pd.read_csv("D:/LTIM/passfail.csv")
# 選取需要的欄位 (BSN、時間、點位、實際X位置、實際Y位置、實際膠長、膠寬、面積、passfail...)
df = data.iloc[:,[0,1,2,3,16,17,18,19,20,21,29,30,31]]
# 資料隨時間排序
df = df.sort_values(["時間"],ascending=True)
# 畫圖 a=資料, b=分類欄位, c=x軸, d=y軸
def passfail_plot(a, b, c, d):
# 畫pass跟fail的散佈圖
plt.figure(figsize=(10,5))
plt.xlabel("Position_x", fontweight = "bold")
plt.ylabel("Area(mm)", fontweight = "bold")
plt.title("Scatter of pass fail in position_X and area(mm)", fontsize = 15, fontweight = "bold")
sns.scatterplot(x=a.iloc[:,c], y=a.iloc[:,d], hue=b)
plt.legend(fontsize=10)
plt.show()
# 畫pass跟fail的圓餅圖
plt.figure( figsize=(10,5) )
b.value_counts().plot( kind='pie', colors=['lightcoral','skyblue'], autopct='%1.2f%%' )
plt.title( 'Pass/Fail data' )
plt.ylabel( '' )
plt.show()
#============================================上蓋=====================================================
# %%
# 選取上蓋資料
top_all = df[(df["位置"] == "Top")]
#%%
passfail_plot(top_all, top_all.iloc[:,-3], 4, -4)
#%%
# 資料分割 .8訓練集.2測試集,以x,y位置、面積...做特徵
train_x,test_x,train_y,test_y = train_test_split(top_all.iloc[:,[4,5,6,7,9]], top_all.iloc[:,-3], test_size=0.2, random_state=1)
# 資料不平等
# oversampling
X_re, y_re = SMOTE(random_state=42).fit_resample(train_x, train_y)
passfail_plot(X_re, y_re, 1, -1)
# undersampling
nearmiss = NearMiss (version=3)
X_train_nearmiss, y_train_nearmiss= nearmiss.fit_resample(train_x, train_y)
passfail_plot(X_train_nearmiss, y_train_nearmiss, 1, -1)
#%%
# 邏輯式迴歸模型
lr_over = LogisticRegression().fit(X_re, y_re)
y_pred_over = lr_over.predict(test_x)
print(classification_report(test_y, y_pred_over))
tn, fp, fn, tp = confusion_matrix(test_y, y_pred_over).ravel()
print(tn, fp, fn, tp)
lr_under = LogisticRegression().fit(X_train_nearmiss, y_train_nearmiss)
y_pred_under = lr_under.predict(test_x)
print(classification_report(test_y, y_pred_under))
tn, fp, fn, tp = confusion_matrix(test_y, y_pred_under).ravel()
print(tn, fp, fn, tp)
# %%
# SVM模型
# over
clf_over=svm.SVC(kernel='rbf',C=1,gamma='auto')
clf_over.fit(X_re, y_re)
clf_pred_over = clf_over.predict(test_x)
print(classification_report(test_y, clf_pred_over))
tn, fp, fn, tp = confusion_matrix(test_y, clf_pred_over).ravel()
print(tn, fp, fn, tp)
# under
clf_under=svm.SVC(kernel='rbf',C=1,gamma='auto')
clf_under.fit(X_train_nearmiss, y_train_nearmiss)
clf_pred_under = clf_under.predict(test_x)
print(classification_report(test_y, clf_pred_under))
tn, fp, fn, tp = confusion_matrix(test_y, clf_pred_under).ravel()
print(tn, fp, fn, tp)
##======================================下蓋===============================================
# %%
bot_all = df[(df["位置"] == "Bot")]
#重排索引值
bot_all.reset_index(inplace=True,drop=True)
# 將品質不良原因判斷欄位不是pass的資料作為fail資料
for i in range(len(bot_all)):
if (bot_all.iloc[i,-2] != "pass"):
bot_all.iloc[i,-2]="fail"
passfail_plot(bot_all, bot_all.iloc[:,-2], 4, -4)
#%%
# 資料分割 .8訓練集.2測試集,以x,y位置、面積...做特徵
train_x,test_x,train_y,test_y = train_test_split(bot_all.iloc[:,[4,5,6,7,9]], bot_all.iloc[:,-2], test_size=0.2, random_state=1)
# 資料不平等,採樣方式:
# oversampling
X_re, y_re = SMOTE(random_state=42).fit_resample(train_x, train_y)
passfail_plot(X_re, y_re, 1, -1)
# undersampling
nearmiss = NearMiss (version=3)
X_train_nearmiss, y_train_nearmiss= nearmiss.fit_resample(train_x, train_y)
passfail_plot(X_train_nearmiss, y_train_nearmiss, 1, -1)
# %%
# 邏輯式迴歸模型
lr_over = LogisticRegression().fit(X_re, y_re)
y_pred_over = lr_over.predict(test_x)
print(classification_report(test_y, y_pred_over))
lr_under = LogisticRegression().fit(X_train_nearmiss, y_train_nearmiss)
y_pred_under = lr_under.predict(test_x)
print(classification_report(test_y, y_pred_under))
# %%
# SVM模型
# over 要跑一段時間!!
clf_over=svm.SVC(kernel='rbf',C=1,gamma='auto')
clf_over.fit(X_re, y_re)
clf_pred_over = clf_over.predict(test_x)
print(classification_report(test_y, clf_pred_over))
tn, fp, fn, tp = confusion_matrix(test_y, clf_pred_over).ravel()
print(tn, fp, fn, tp)
#%%
# under
clf_under=svm.SVC(kernel='rbf',C=1,gamma='auto')
clf_under.fit(X_train_nearmiss, y_train_nearmiss)
clf_pred_under = clf_under.predict(test_x)
print(classification_report(test_y, clf_pred_under))
tn, fp, fn, tp = confusion_matrix(test_y, clf_pred_under).ravel()
print(tn, fp, fn, tp)
# %%
# 檢查錯誤判斷的fail資料中,有幾筆是品質檢測結果真的fail的資料
ypd = pd.DataFrame(test_y)
ypd['index'] = ypd.index
indexposition=[]
for i in range(len(test_y)):
if ypd.iloc[i,0] != clf_pred_over[i]:
indexposition.append(ypd.iloc[i,1])
real_fail = bot_all.iloc[indexposition, -3]
len(real_fail[real_fail=="fail"])/len(bot_all[bot_all.iloc[:,-3]=='fail']) # 所有真實fail資料中誤判的比例