def satWaterPressure(T, unit='kelvin'):
"""
Returns the saturated water vapor pressure according eq (3.97) of Wallace and Hobbes, page 99.
e0, b, T1 and T2 are constants specific for water vapor
Parameters
----------
T: float
thermodynamic temperature
Returns
-------
saturated vapor pressure of water (in kPa)
"""
from np import exp
e0=0.61094
b=17.2694
if unit=='kelvin':
T1=273.16
T2=35.86
elif units=='celsius':
T1=0.
T2=243.04
else:
raise TypeError('Check your units')
brackets=b*(T-T1)/(T-T2)
return e0*exp(brackets)
def satWaterPressure(T, unit='kelvin'):
"""
Returns the saturated water vapor pressure according eq (3.97) of Wallace and Hobbes, page 99.
e0, b, T1 and T2 are constants specific for water vapor
Parameters
----------
T: float
thermodynamic temperature
Returns
-------
saturated vapor pressure of water (in kPa)
"""
from np import exp
e0 = 0.61094
b = 17.2694
if unit == 'kelvin':
T1 = 273.16
T2 = 35.86
elif units == 'celsius':
T1 = 0.
T2 = 243.04
else:
raise TypeError('Check your units')
brackets = b * (T - T1) / (T - T2)
return e0 * exp(brackets)
def RHS(x, y, z):
return -exp(-(x - 0.25)**2 - (y - 0.6)**2 - z**2)
def softmax(x):
m = np.max(x)
e = np.exp(x - m)
return e / e.sum()
def density(T):
return 8.9975852012753705e3 * np.exp(
-1.0e-6 *
(+13.251 * T + 6.903e-3 / 2.0 * T**2 + 8.5306e-7 / 3.0 * T**3))
def electrical_conductivity(T):
return 1e6 / (28.9 - 18.8 * np.exp(-(np.log(T / 1023.0) / 2.37)**2))
def blurringkernel(shape, T, a, b):
xx, yy = shape
x, y = np.ogrid[(-xx / 2):(xx / 2), (-yy / 2):yy / 2]
q = (np.pi * (x * a + y * b))
q[np.where(q == 0)] = T
return (T / q) * np.sin(q) * np.exp(-1j * q)
def f(x):
return np.exp(-x) - x**2/2
def secDerf(x):
return np.exp(-x) - 1
def derf(x):
return - np.exp(-x) - x
# Python实现正态分布
import np
import matplotlib.pyplot as plt
import math
# 绘制正态分布概率密度函数
u = 0 # 均值μ
u01 = -2
sig = math.sqrt(0.2) # 标准差δ
sig01 = math.sqrt(1)
sig02 = math.sqrt(5)
sig_u01 = math.sqrt(0.5)
x = np.linspace(u - 3*sig, u + 3*sig, 50)
x_01 = np.linspace(u - 6 * sig, u + 6 * sig, 50)
x_02 = np.linspace(u - 10 * sig, u + 10 * sig, 50)
x_u01 = np.linspace(u - 10 * sig, u + 1 * sig, 50)
y_sig = np.exp(-(x - u) ** 2 /(2* sig **2))/(math.sqrt(2*math.pi)*sig)
y_sig01 = np.exp(-(x_01 - u) ** 2 /(2* sig01 **2))/(math.sqrt(2*math.pi)*sig01)
y_sig02 = np.exp(-(x_02 - u) ** 2 / (2 * sig02 ** 2)) / (math.sqrt(2 * math.pi) * sig02)
y_sig_u01 = np.exp(-(x_u01 - u01) ** 2 / (2 * sig_u01 ** 2)) / (math.sqrt(2 * math.pi) * sig_u01)
plt.plot(x, y_sig, "r-", linewidth=2)
plt.plot(x_01, y_sig01, "g-", linewidth=2)
plt.plot(x_02, y_sig02, "b-", linewidth=2)
plt.plot(x_u01, y_sig_u01, "m-", linewidth=2)
# plt.plot(x, y, 'r-', x, y, 'go', linewidth=2,markersize=8)
plt.grid(True)
plt.show()
logProbability0[np.isinf(logProbability0)] = -100 # заменяем -бесконечность
logLikehood0 = logProbability0.sum()
#print(logLikehood0)
r = 1
kde1 = KernelDensity(kernel='gaussian', bandwidth=0.05)
kde1.fit(dataTrain.loc[dataTrain['Response'] == r,
'L1_S24_F1846'].values.reshape(-1, 1))
logProbability1 = kde1.score_samples(
dataTest.loc[dataTest['Response'] == r,
'L1_S24_F1846'].values.reshape(-1, 1))
logProbability1[np.isinf(logProbability1)] = -100 # заменяем -бесконечность
logLikehood1 = logProbability1.sum()
#print(logLikehood1)
predictionProbXafter0 = np.exp(
kde0.score_samples(dataTest['L1_S24_F1846'].values.reshape(-1, 1)))
predictionProbXafter1 = np.exp(
kde1.score_samples(dataTest['L1_S24_F1846'].values.reshape(-1, 1)))
predictionProb0afterX = predictionProbXafter0 #тут должна быть формула Баеса
predictionProb1afterX = predictionProbXafter1 #тут должна быть формула Баеса
ind = np.argsort(
predictionProb0afterX) # сортировка, возвращающая индексы элементов
print(predictionProb0afterX[ind[-10:]]) # вывод последних 10 элементов
print(sum(dataTest.loc[ind[-100:], 'Response'])
) # количество бракованных среди 100 с максимальной вероятностью брака
ind1 = np.argsort(
predictionProb1afterX) # сортировка, возвращающая индексы элементов
print(predictionProb1afterX[ind1[-10:]]) # вывод последних 10 элементов
print(sum(dataTest.loc[ind1[-100:], 'Response'])
from django.test import TestCase
import np
# Create your tests here.
X = np.array([[0, 0, 1], [0, 1, 1], [1, 0, 1], [1, 1, 1]])
y = np.array([[0, 1, 1, 0]]).T
syn0 = 2 * np.random.random((3, 4)) - 1
syn1 = 2 * np.random.random((4, 1)) - 1
for j in xrange(60000):
l1 = 1 / (1 + np.exp(-(np.dot(X, syn0))))
l2 = 1 / (1 + np.exp(-(np.dot(l1, syn1))))
l2_delta = (y - l2) * (l2 * (1 - l2))
l1_delta = l2_delta.dot(syn1.T) * (l1 * (1 - l1))
syn1 += l1.T.dot(l2_delta)
syn0 += X.T.dot(l1_delta)
def gaussian_kernel_1d(sigma):
kernel_radius = np.ceil(sigma) * 3
kernel_size = kernel_radius * 2 + 1
ax = np.arange(-kernel_radius, kernel_radius + 1., dtype=np.float32)
kernel = np.exp(-(ax**2) / (2. * sigma**2))
return (kernel / np.sum(kernel)).reshape(1, kernel.shape[0])
def sigmoid(x):
# 應用sigmoid啟用函式
return 1 / (1 + np.exp(-x))
def boltzman_probability(loss1, loss2, temp):
k = 1.
return np.exp((loss1 - loss2) / (k * temp))
