def plotmhd(file_DRR_AP, title=None, margin=0.0, dpi=300):
#plot the mhd DRR and return the ax and plt
itkimage = sitk.ReadImage(file_DRR_AP)
nda = sitk.GetArrayFromImage(itkimage)
spacing = itkimage.GetSpacing()
xsize = nda.shape[1]
ysize = nda.shape[0]
figsize = (1 + margin) * xsize / dpi, (1 + margin) * ysize / dpi
fig = plt.figure(figsize=figsize, dpi=dpi)
#print "image pixel numbers", xsize, ysize dpi:dots per inch
#print "figure size", figsize
#ax = fig.add_subplot(111)
ax = plt.Axes(fig, [0, 0, 1, 1], frameon=False)
fig.add_axes(ax)
extent = (0, xsize * spacing[0], ysize * spacing[1], 0
) # the same orientation as DRR
#print "extent", extent
t = ax.imshow(nda, extent=extent, interpolation=None)
if nda.ndim == 2:
t.set_cmap("gray")
if (title):
plt.tile(title)
#plt.show()
return ax, fig
def plot_cross_validation(self,
figsize=(6, 4),
figname="metrics.png",
show=True):
"""
Plots Cross validation
Input:
- figsize: tuple, the size of the metric curve
- figname: string, what you want the combined plot to be saved as
- show: boolean, whether you want to plt.show() your figure or just save it to your computer
"""
file_name = "K_fold_Cross_Validation.png"
# check_for_array(self.cross_val, "Cross_Val")
plt.figure(figsize=figsize)
plt.tile("K-fold Cross Validation", fontsize=14)
plt.yticks(fontsize=14)
plt.xticks(fontsize=14)
plt.ylabel("Folds", fontsize=14)
plt.xlabel("Accuracy", fontsize=14)
# plt.plot(self.loss)
plt.plot(self.cross_val)
plt.savefig(file_name)
if show: plt.show()
plt.close()
plt.savefig(out+'/yq_tot_hist.png', dpi=190)
if False:
out='/home/dw1519/dw1519/galex/plots/co239-10/radec'
co_list = []
info_list = []
for i in range(0, 28):
co = np.load(out+'/{0}.npy'.format((i+1)*100))/36000./800/0.001666*2400
info = np.load(out+'/info_{0}.npy'.format((i+1)*100))[1:]
co_list.append(co)
info_list.append(info)
co = np.concatenate(co_list, axis=0)
info = np.concatenate(info_list, axis=0)
plt.hist(info[:,1],32,alpha=0.5)
plt.tile('{0}, photons'.format(32, np.sum(info[:,1]<2)/np.sum(info[:,1]>=0)))
plt.savefig(out+'/ya_tot_hist.png', dpi=190)
if False:
scan = '2813'
name = 'AIS_GAL_SCAN_0'+scan+'_0003'
out = '../data/'+name+'-cal-sec'
data = np.load('../data/'+name+'-cal-sec/photon_match_d_081017.npy')
print scan
data_mask = data[:,-3]>-1
data = data[data_mask]
dy = data[:,-1]/36000./800/0.001666*2400
dx = data[:,-2]/36000./800/0.001666*2400
ya = data[:,-5]
q = data[:,-6]
ya = data[:,-7]
def save_plot(self,circuit,i,log=True,name='',**kwargs):
v = circuit.evaluate(self.spice_commands[i])
#For every measurement in results
for k in v.keys():
score = self._rank(v,i,k)
plt.figure()
freq = v[k][0]
gain = v[k][1]
goal_val = [self.ff[i](c,k) for c in freq]
if self.plot_weight:
weight_val = [self.fitness_weight[i](c,k) for c in freq]
if self.constraints[i]!=None and self.plot_constraints:
constraint_val = [not self.constraints[i](freq[c],gain[c],k) for c in xrange(len(freq))]
if log==True:#Logarithmic plot
plt.semilogx(freq,gain,'g',basex=10)
plt.semilogx(freq,goal_val,'b',basex=10)
if self.plot_weight:
plt.semilogx(freq,weight_val,'r--',basex=10)
if self.plot_constraints:
plt.semilogx(freq,constraint_val,'m',basex=10)
else:
plt.plot(freq,gain,'g')
plt.plot(freq,goal_val,'b')
if self.plot_weight:
plt.plot(freq,weight_val,'r--')
if self.constraints[i]!=None and self.plot_constraints:
plt.plot(freq,constraint_val,'m')
# update axis ranges
ax = []
ax[0:4] = plt.axis()
# check if we were given a frequency range for the plot
if k in self.plot_yrange.keys():
plt.axis([min(freq),max(freq),self.plot_yrange[k][0],self.plot_yrange[k][1]])
else:
plt.axis([min(freq),max(freq),min(-0.5,-0.5+min(goal_val)),max(1.5,0.5+max(goal_val))])
if self.sim_type[i]=='dc':
plt.xlabel("Input (V)")
if self.sim_type[i]=='ac':
plt.xlabel("Input (Hz)")
if self.sim_type[i]=='tran':
plt.xlabel("Time (s)")
if self.plot_titles!=None:
try:
plt.title(self.plot_titles[i][k])
except:
plt.tile(k)
plt.annotate('Generation '+str(self.generation),xy=(0.05,0.95),xycoords='figure fraction')
if score!=None:
plt.annotate('Score '+'{0:.2f}'.format(score),xy=(0.75,0.95),xycoords='figure fraction')
plt.grid(True)
# turn on the minor gridlines to give that awesome log-scaled look
plt.grid(True,which='minor')
if k[0]=='v':
plt.ylabel("Output (V)")
elif k[0]=='i':
plt.ylabel("Output (A)")
plt.savefig('plot'+strftime("%Y-%m-%d %H:%M:%S")+'-'+k+'-'+name+'.png')
if False:
out = '/home/dw1519/dw1519/galex/plots/co239-10/radec'
co_list = []
info_list = []
for i in range(0, 28):
co = np.load(out + '/{0}.npy'.format(
(i + 1) * 100)) / 36000. / 800 / 0.001666 * 2400
info = np.load(out + '/info_{0}.npy'.format((i + 1) * 100))[1:]
co_list.append(co)
info_list.append(info)
co = np.concatenate(co_list, axis=0)
info = np.concatenate(info_list, axis=0)
plt.hist(info[:, 1], 32, alpha=0.5)
plt.tile('{0}, photons'.format(
32,
np.sum(info[:, 1] < 2) / np.sum(info[:, 1] >= 0)))
plt.savefig(out + '/ya_tot_hist.png', dpi=190)
if True:
scan = '0815'
name = 'AIS_GAL_SCAN_0' + scan + '_0001'
out = '../data/' + name + '-cal-sec'
data = np.load('../data/' + name +
'-cal-sec/photon_match_d_072217.npy')
print scan
dy = data[:, -1] / 36000. / 800 / 0.001666 * 2400
dx = data[:, -2] / 36000. / 800 / 0.001666 * 2400
q = data[:, -6]
ya = data[:, -7]
r2_score(y_train, model.predict(x_train))
model.score(x_train, y_train)
model.score(x_test, y_test)
'''
= 0.5171.. => 낮은 예측률
따라서 리그레이션 말고 다른 모델로 예측하기 추천
'''
plt.scatter(y_train, model.predict(x_train))
plt.xlabel('true values')
plt.ylabel('predictions')
plt.title('train')
plt.scatter(y_test, pred)
plt.xlabel('true values')
plt.ylabel('predictions')
plt.tile('test')
#cross validation
from sklearn.cross_validation import cross_val_score, cross_val_predict
scores = cross_val_score(model, x_train, y_train, cv=6)
pred = cross_val_predict(model, x_train, y_train, cv=6)
plt.scatter(y, pred)
#k fold cross validation
import numpy as np
from sklearn.model_selection import KFold
x = np.array([[1, 2], [3, 4], [1, 2], [3, 4]])
y = np.array([1, 2, 3, 4])
kf = KFold(n_splits=2)
kf.get_n_splits(x)
import matplotlib.pyplot as plt
import pandas as pd
file = pd.read_csv("C:/Users/gestor/Downloads/a2_MANCHAS(1).csv")
x = list(file["Ano"])
y = list(file["manchas"])
plt.plot(x, y)
plt.tile("Demonstrando as manchas solares de ano em ano(30 primeiros)",
fontsize=15,
color="r",
style="italic",
family="monospace")
plt.title("Demonstrando as manchas solares de ano em ano(30 primeiros)",
fontsize=15,
color="r",
style="italic",
family="monospace")
plt.xlabel("Ano", size=12)
plt.ylabel("Manchas", size=12)
plt.grid()
plt.tight_layout()
plt.savefig("30first.jpg")
plt.savefig("30first.png")
### Covariances
cov = []
for i in range(20, 66):
tryt = g.loc[g['age'] == i].cov()
tryt = np.array(tryt)
cov.append(tryt[0:3, 0:3])
del tryt
cov_ci = np.zeros((46, 1))
for j in range(0, 46):
cov_ci[j, 0] = cov[j][0, 1]
plt.plot(ages, cov_ci)
plt.tile('Lifecycile CI Covariance')
del cov_ci
cov_cw = np.zeros((46, 1))
for j in range(0, 46):
cov_cw[j, 0] = cov[j][0, 2]
plt.plot(ages, cov_cw)
plt.tile('Lifecycile CW Covariance')
del cov_cw
cov_wi = np.zeros((46, 1))
for j in range(0, 46):
cov_wi[j, 0] = cov[j][1, 2]
plt.plot(ages, cov_wi)
plt.title('PCA Variance across each dimension')
pca_top2 = pca.transform(x)[:, :2]
plt.figure()
plt.scatter(pca_top2[:, 0], pca_top2[:, 1], s=0.01)
plt.title('PCA Top-2 Dimensions')
from sklearn.manifold import TSNE
tsne = TSNE(n_components=2, perplexity=30, learning_rate=100)
tsne_dims2 = tsne.fit_transform(x)
tsne_c = TSNE(n_components=2, perplexity=30, learning_rate=100)
tsne_dims2_c = tsne_c.fit_transform(x_c)
plt.figure()
plt.scatter(tsne_dims2[:, 0], tsne_dims2[:, 1], s=0.1, c='r')
#plt.title('Clean_Data: TSNE (down-to) 2 Dimensions');
plt.scatter(tsne_dims2_c[:, 0], tsne_dims2_c[:, 1], s=0.1, c='b')
#plt.title('Corrupt_data: TSNE (down-to) 2 Dimensions');
plt.tile('tSNE')
plt.show()
#%% 6. FFT on the time-series data
fft_x = np.abs(np.fft.fft(x, axis=1))
fft_x = fft_x[:, :int(np.ceil(x.shape[1] / 2)) + 1]
for label in range(y.shape[-1]):
labelled = fft_x[np.where(y[:, label] == 1)[0], :]
fig_rows, fig_cols = 5, 2
fig, ax = plt.subplots(fig_rows, fig_cols, figsize=(16, 5))
row_inds = np.random.choice(labelled.shape[0],
fig_rows * fig_cols,
replace=False)
for fig_row in range(fig_rows):
for fig_col in range(fig_cols):
ax[fig_row, fig_col].plot(
#rf = np.fft.fft(r[iworm, :ntimes]);
#rf = 2.0/ntimes * np.abs(rf[:ntimes//2]);
#ff = np.linspace(0.0, 1.0/(2.0*dt), ntimes//2)
### FFT
ff, pfsd_r, = signal.periodogram(d[iworm, :ntimes], 1 / dt)
nmean = 50
pfsd_r_m = np.convolve(pfsd_r, np.ones(nmean) / nmean, mode='same')
plt.figure(1 * fi)
plt.clf()
plt.semilogy(ff, pfsd_r)
plt.semilogy(ff, pfsd_r_m)
plt.tile('power %s %s %d' % (strain, fn, iworm))
### Spectogram
f, t, Sxx = signal.spectrogram(d[iworm, :ntimes],
1 / dt,
nperseg=2**14,
noverlap=2**14 - 2**8)
plt.figure(2 * fi)
plt.clf()
ax = plt.subplot(2, 1, 1)
nfmax = 50
plt.pcolormesh(t, f[:nfmax], Sxx[:nfmax, :])
plt.ylabel('Frequency [Hz]')
plt.xlabel('Time [sec]')
plt.tile('spectogram %s %s %d' % (strain, fn, iworm))
"""
h1 = np.arange(-1, 1, res1)
gauss = intensity(wi, lamda, P, -1, 1, res1)
"""
BEAM NEL FUOCO
"""
h2 = np.arange(-0.01, 0.01, res2)
gauss_after_lens = intensity(w1, lamda, P, -0.01, 0.01, res2)
"""
BEAM DOPO LA SECONDA LENTE
"""
h3 = np.arange(-10, 10, res3)
gauss_after_second_lens = intensity(w2, lamda, P, -10, 10, res3)
plt.figure(1)
plt.tile('Beam before the SF')
plt.plot(h1, gauss, color="Red")
plt.xlabel("Contour Radius (mm)")
plt.ylabel("Percent Irradiance")
plt.show()
plt.clf()
plt.figure(2)
plt.tile('Beam in the focal point (pinhole position)')
plt.plot(h2, gauss_after_lens, color="Red")
plt.xlabel("Contour Radius (mm)")
plt.ylabel("Percent Irradiance")
plt.xticks(np.arange(-0.01, 0.015, step=0.005))
plt.show()
plt.clf()
# Building the Model
model = tf.keras.Sequential([
tf.keras.layers.Embedding(tokenizer.vocab_size, 64),
tf.keras.layers.Conv1D(128, 5, activation='relu'),
tf.keras.layers.GlobalAveragePooling1D(),
tf.keras.layers.Dense(64, activation='relu'),
tf.keras.layers.Dense(1, activation='sigmoid')
])
print(model.summary())
# Compile the model
model.compile(optimizer='adam',
loss='binary_crossentropy',
metrics=['accuracy'])
# Train the model
NUM_EPOCHS = 1 # Traning only on single Epoch, as it takes a long time to train. Try with more Epochs!
history = model.fit(train_dataset,
epochs=NUM_EPOCHS,
validation_data=test_dataset)
print(history.history.keys())
plt.plot(history.history['acc'])
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.tile('Model Training')
plt.show()
cummulate_survival_ratio = []
for i in range(1, 80):
cummulate_survival_ratio.append(
df_train[df_train["Age"] < i]["Survived"].mean())
pd.DataFrame([0, pd.NA, 1]).mean()
cummulate_survival_ratio = []
for i in range(1, 80):
cummulate_survival_ratio.append(
df_train["Survived"][df_train["Age"] < i].mean())
plt.figure(figsize=(7, 7))
plt.plot(cummulate_survival_ratio)
plt.tile("Survival rate change depending on range of Age", y=1.02)
plt.ylabel("Survival rate")
plt.xlabel("Range of Age(0~x)")
plt.show()
f, ax = plt.subplots(1, 2, figsize=(18, 8))
sns.violinplot("Pclass",
"Age",
hue="Survived",
data=df_train,
scale="count",
split=True,
ax=ax[0])
ax[0].set_title("Pclass and Age vs Survived")
ax[0].set_yticks(range(0, 110, 10))
sns.violinplot("Sex",
sizes = [15, 30, 45, 10] # 占比
explode = (0, 0.1, 0, 0)
plt.pie(sizes, explode=explode, labels=labels,
autopct='%1.1f%%',shadow=False, startangle=90) # 饼图
plt.axis('equal')
# 例2 直方图：
np.random.seed(0)
mu, sigma = 100, 20
a = np.random.normal(mu, sigma, size=100)
plt.hist(a,
20, # 直方图的个数
normed=1, histtype='stepfilled',
facecolor='b', alpha=0.75)
plt.tile('Histogram')
plt.show()
# 例3：散点图 面向对象的方式绘制
fig, ax = plt.subplots() # 将subplots 函数变成object，分别对应函数生成的图表和图表对应的区域；
# 为空时，默认为111，对应的绘图区域（即当前的绘图区域）是ax
ax.plot(10*np.random.randn(100), 10*np.random.randn(100), 'o') # 在ax(绘图区）中绘制
ax.set_title('Simple Scatter')
# 使用ax这种面向对象的方法，所有.plot函数和标题设置函数 变成了object的methods，而不再是plt下面的函数。官方推荐该方法。
plt.show()
# √ 横轴是日期
plt.plot_date(date,data)
# √ 将 x轴 y轴改为对数标度
pylab.semilogx()
