-
Notifications
You must be signed in to change notification settings - Fork 0
/
p4.py
141 lines (103 loc) · 3.81 KB
/
p4.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
from data import loadData
from sklearn import linear_model
from sklearn.svm import SVC
import pandas as pd
import numpy as np
from sklearn.model_selection import KFold
from sklearn.multiclass import OneVsRestClassifier
import matplotlib.pyplot as plt
from sklearn.metrics import precision_recall_curve
from sklearn.metrics import roc_curve, auc
data = loadData()
df_x = data.loc[:,'b3356':'b4703']
df_y = data.loc[:,'GrowthRate']
Lasso_model = linear_model.Lasso(alpha = 1e-5,tol = 0.001)
Lasso_model.fit(df_x,df_y)
coef = Lasso_model.coef_
print('Features with nonzero weights are used:')
print('Number of features Used For all Models:',len(coef != 0))
print()
df_x = df_x.loc[:,coef != 0] #features that with nonzero weight
y1 = data.loc[:,'Strain']
y2 = data.loc[:,'Medium']
y3 = data.loc[:,'Stress']
y4 = data.loc[:,'GenePerturbed']
def p4PR(x,y):
df_x = x
df_y1 = y
dummies_y1 = pd.get_dummies(df_y1)
num_fold = 10
k_fold = KFold(n_splits=num_fold, shuffle=True, random_state=12345)
model1 = OneVsRestClassifier(SVC(probability=True))
y_real = []
y_pred = []
for i, (trainIndex, testIndex) in enumerate(k_fold.split(df_x)):
xtrain,xtest = df_x.iloc[trainIndex,:],df_x.iloc[testIndex,:]
ytrain,ytest = dummies_y1.iloc[trainIndex,:],dummies_y1.iloc[testIndex,:]
model1.fit(xtrain,ytrain)
y_predict = model1.predict_proba(xtest)
y_pred.append(y_predict)
y_real.append(ytest) #appending ytest data as real data set
##For PR Curve
precision, recall, _ = precision_recall_curve(np.array(ytest).ravel(), y_predict.ravel())
pr_auc = auc(recall,precision)
plt.plot(recall,precision)
print('Precision Recall Curve:')
print('Fold: {} - AUC: {}'.format(i+1, pr_auc))
y_real = np.concatenate(y_real)
y_pred = np.concatenate(y_pred)
precision, recall, _ = precision_recall_curve(y_real.ravel(), y_pred.ravel())
auc_overall = auc(recall,precision)
print('Overall AUC for PR:',auc_overall)
print()
plt.plot(recall,precision,color='black')
plt.xlabel('recall')
plt.ylabel('precision')
plt.legend(['Fold1','Fold2','Fold3','Fold4','Fold5','Fold6','Fold7','Fold8','Fold9','Fold10','Overall'],loc='lower center')
plt.show()
def p4ROC(x,y):
df_x = x
df_y1 = y
dummies_y1 = pd.get_dummies(df_y1)
num_fold = 10
k_fold = KFold(n_splits=num_fold, shuffle=True, random_state=12345)
model1 = OneVsRestClassifier(SVC(probability=True))
y_real = []
y_pred = []
for i, (trainIndex, testIndex) in enumerate(k_fold.split(df_x)):
xtrain,xtest = df_x.iloc[trainIndex,:],df_x.iloc[testIndex,:]
ytrain,ytest = dummies_y1.iloc[trainIndex,:],dummies_y1.iloc[testIndex,:]
model1.fit(xtrain,ytrain)
y_predict = model1.predict_proba(xtest)
y_pred.append(y_predict)
y_real.append(ytest) #appending ytest data as real data set
#For ROC Curve
fp_rate, tp_rate, _ = roc_curve(np.array(ytest).ravel(), y_predict.ravel())
roc_auc = auc(fp_rate, tp_rate)
plt.plot(fp_rate, tp_rate)
print('ROC Curve')
print('Fold: {} - AUC: {}'.format(i+1, roc_auc))
y_real = np.concatenate(y_real)
y_pred = np.concatenate(y_pred)
fp_rate, tp_rate, _ = roc_curve(np.array(y_real).ravel(), y_pred.ravel())
auc_overall = auc(fp_rate,tp_rate)
print('Overall AUC for ROC:',auc_overall)
print()
plt.plot(fp_rate,tp_rate,color='black')
plt.xlabel('False Positive Rate')
plt.ylabel('True Positive Rate')
plt.legend(['Fold1','Fold2','Fold3','Fold4','Fold5','Fold6','Fold7','Fold8','Fold9','Fold10','Overall'],loc='lower center')
plt.show()
#
# p4PR(df_x,y1)
# p4ROC(df_x,y1)
#
# p4PR(df_x,y2)
# p4ROC(df_x,y2)
#
# p4PR(df_x,y3)
# p4ROC(df_x,y3)
#
# p4PR(df_x,y4)
# p4ROC(df_x,y4)
#