-
Notifications
You must be signed in to change notification settings - Fork 0
/
DNN_plot.py
156 lines (134 loc) · 5.98 KB
/
DNN_plot.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
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.colors as colors
from sklearn.preprocessing import normalize
from sklearn.metrics import roc_curve, auc, f1_score
from sklearn.metrics import confusion_matrix,accuracy_score,average_precision_score,roc_auc_score
def outputWeights(model, datadir):
weights = model.layers[0].get_weights()[0]
f = open(datadir+"/weights_layer1.txt")
f.write(str(weights))
print("Event weights: "+str(weights))
return weights
def plot_accuracy(history, outputdir):
fig_acc = plt.figure(figsize=(12,8))
plt.plot(history.history['accuracy'],color="blue")
plt.plot(history.history['val_accuracy'],color="red")
plt.title('',fontsize=20)
plt.ylabel('weighted accuracy',fontsize=30, labelpad=15)
plt.xlabel('epoch',fontsize=30, labelpad=15)
plt.xticks(fontsize=25)
plt.yticks(fontsize=25)
plt.legend(['training', 'validation'], loc='lower right',fontsize=30)
plt.tight_layout()
fig_acc.savefig(outputdir+'/plots/accuracy.png')
def plot_accuracy_partial(history, outputdir):
fig_acc = plt.figure(figsize=(12,8))
plt.plot(history.history['acc_sig'],color="blue")
plt.plot(history.history['val_acc_sig'],color="red")
plt.title('',fontsize=20)
plt.ylabel('weighted accuracy',fontsize=20)
plt.xlabel('epoch',fontsize=20)
plt.legend(['train', 'validation'], loc='lower right',fontsize=20)
fig_acc.savefig(outputdir+'/plots/accuracy_sig.png')
fig_acc = plt.figure(figsize=(12,8))
plt.plot(history.history['acc_bkg'],color="blue")
plt.plot(history.history['val_acc_bkg'],color="red")
plt.title('',fontsize=20)
plt.ylabel('weighted accuracy',fontsize=20)
plt.xlabel('epoch',fontsize=20)
plt.legend(['train', 'validation'], loc='lower right',fontsize=20)
fig_acc.savefig(outputdir+'/plots/accuracy_bkg.png')
def plot_loss(history, outputdir):
fig_loss = plt.figure(figsize=(12,8))
plt.plot(history.history['loss'],color="blue")
plt.plot(history.history['val_loss'],color="red")
plt.title('',fontsize=20)
plt.ylabel('loss',fontsize=30,labelpad=15)
plt.xlabel('epoch',fontsize=30,labelpad=15)
plt.xticks(fontsize=25)
plt.yticks(fontsize=25)
plt.legend(['training', 'validation'], loc='upper right',fontsize=30)
plt.tight_layout()
fig_loss.savefig(outputdir+'/plots/loss.png')
def plot_loss_partial(history, outputdir):
fig_acc = plt.figure(figsize=(12,8))
plt.plot(history.history['loss'],color="black")
plt.plot(history.history['nominalLoss'],color="red")
plt.plot(history.history['DisCoLoss'],color="green")
plt.title('',fontsize=20)
plt.ylabel('loss',fontsize=30, labelpad=15)
plt.xlabel('epoch',fontsize=30, labelpad=15)
plt.xticks(fontsize=25)
plt.yticks(fontsize=25)
plt.legend(['total','BCE', 'DisCo'], loc='lower right',fontsize=30)
plt.tight_layout()
fig_acc.savefig(outputdir+'/plots/loss_partial.png')
def plot_output(predictions, labels, outputdir, name="test"):
# Plotting outputs
predictionsSig = []
predictionsBkg = []
for i in range(0,len(labels)):
if(labels[i]==1):
predictionsSig.append(predictions[i])
if(labels[i]==0):
predictionsBkg.append(predictions[i])
predictionsSigPlot = np.asarray(predictionsSig)
predictionsBkgPlot = np.asarray(predictionsBkg)
fout = plt.figure(figsize=(12,8))
plt.hist(predictionsBkgPlot, label='Background', histtype="stepfilled", color="blue", density=True, alpha=0.5)
plt.hist(predictionsSigPlot, label='Signal', histtype="stepfilled", color="red", density=True, alpha=0.5)
plt.title("")
plt.xlabel("DNN output",fontsize=20)
plt.ylabel("Number of events",fontsize=20)
#plt.set_yscale('log')
plt.legend(loc='upper left')
fout.savefig(outputdir+'/plots/output_'+name+'.png')
def plot_ROC(predictions, labels, names, outputdir):
fig_roc=plt.figure()
for i in range(0, len(predictions)):
prediction = predictions[i]
label = labels[i]
name = names[i]
fpr, tpr, _ = roc_curve(label, prediction)
roc_auc = auc(fpr, tpr)
plt.plot(fpr, tpr, label=name+' ROC (area = %0.2f)' % roc_auc)
plt.plot([0, 1], [0, 1], 'k--')
plt.xlim([0.0, 1.05])
plt.ylim([0.0, 1.05])
plt.xlabel('False Positive Rate',fontsize=15)
plt.ylabel('True Positive Rate', fontsize=15)
plt.legend(loc='lower right')
print('AUC: %f' % roc_auc)
fig_roc.savefig(outputdir+'/plots/ROC.png')
def plot_2D(output, HT, weights, outputdir):
output = np.ravel(output)
assert(len(output) == len(HT))
H, xedges, yedges = np.histogram2d(output, HT, range=[[0,1],[0,3000]], weights=weights)
H = H.T
H_norm = normalize(H, norm="l1")
fig1=plt.figure()
mesh1 = plt.pcolormesh(xedges, yedges, H, norm=colors.LogNorm(vmin=H[H > 0].min(), vmax=H.max()))
fig1.colorbar(mesh1)
fig1.savefig(outputdir+'/plots/2D.png')
fig=plt.figure()
mesh = plt.pcolormesh(xedges, yedges, H_norm, norm=colors.LogNorm(vmin=H_norm[H_norm > 0].min(), vmax=H_norm.max()))
fig.colorbar(mesh)
fig.savefig(outputdir+'/plots/2Dnormed.png')
def plot_ST(STsig, weightsig, STbkg, weightbkg, outputdir):
fig=plt.figure()
plt.hist(STsig, range(0, 6000, 100), weights=weightsig, label="Signal", alpha = 0.5, log=True)
plt.hist(STbkg, range(0, 6000, 100), weights=weightbkg, label="Background", alpha = 0.5, log=True)
plt.legend()
fig.savefig(outputdir+'/plots/ST.png')
def plot_hist(data, label, outputdir, weights=np.zeros(1)):
fig=plt.figure()
if(np.array_equal(weights, np.zeros(1))): weights=np.ones(len(data))
plt.hist(data, np.arange(0.95*np.amin(data), 1.05*np.amax(data), (np.amax(data)-np.amin(data))/100), weights=weights, label=label, log=True)
plt.legend()
fig.savefig(outputdir+'/plots/'+label+'.png')
def doHistoryPlots(history, outputdir):
plot_accuracy(history, outputdir)
plot_accuracy_partial(history, outputdir)
plot_loss(history, outputdir)
plot_loss_partial(history, outputdir)