def np_logistic_distribution():
x = random.logistic(loc=1, scale=2, size=(2, 3))
print(x)
# sns.distplot(random.logistic(size=1000), hist=False)
sns.distplot(random.normal(scale=2, size=1000), hist=False, label='normal')
sns.distplot(random.logistic(size=1000), hist=False, label='logistic')
plt.show()
def synthetic_data():
mixing_0 = empirical_mn([1, 1, 1], eye(3), size=500)
mixing_1 = empirical_mn([-1, -1, -1], eye(3), size=500)
sources = logistic(0, 1, size=(3, 50000))
data = dot(vstack((mixing_0,mixing_1)), sources) +\
uniform(low=1, high=2, size=(1, 50000))
labels = [0] * 500 + [1] * 500
return (data, labels)
def synthetic_data():
mixing_0 = empirical_mn([1,1,1], eye(3), size = 500)
mixing_1 = empirical_mn([-1,-1,-1], eye(3), size = 500)
sources = logistic(0,1, size=(3,50000))
data = dot(vstack((mixing_0,mixing_1)), sources) +\
uniform(low=1, high=2, size=(1, 50000))
labels = [0]*500 + [1]*500
return (data,labels)
def time_to_mutation_rate(tree):
if not hasattr(GC, "NUMPY_SEEDED"):
from numpy.random import seed as numpy_seed
numpy_seed(seed=GC.random_number_seed)
GC.random_number_seed += 1
GC.NUMPY_SEEDED = True
t = read_tree_newick(tree)
for node in t.traverse_preorder():
if node.edge_length is not None:
node.edge_length *= logistic(loc=GC.tree_rate_loc,
scale=GC.tree_rate_scale)
return str(t)
def logistic_regression_simulation_ystars(nobs,
xs,
param_matrix,
ymeans,
yvars=None):
"""
Method 3: underlying latent variable approach.
Can specify a conditional mean for y here
beta should NOT contain an intercept
Y_i* = beta*x + e where e ~ logistic(b0, 1)
"""
y_stars = xs.dot(param_matrix) + random.logistic(ymeans, 1, nobs).reshape(
nobs, 1)
ys = vint(y_stars > 0.)
return ys
def getImpliedPotOdds(self):
impliedPot = self.potSize
for action in self.actions:
if "CALL" in action:
callSize = int(action.split(':')[1])
elif "RAISE" in action or "BET" in action:
minRaise = int(action.split(':')[1])
maxRaise = int(action.split(':')[2])+minRaise #if we raise
enumRaises = list(range(minRaise,maxRaise+1))
try:
#enumRaises = [0]+[callSize]+enumRaises fold,call,raise
#we shift 0->callSize-1 because statistics
enumRaises = [callSize-1]+[callSize]+enumRaises
except:
pass
try:
if len(enumRaises)>1: #not just callSize-1
expectedActions = dict()
for opponent in self.opponents:
currOpp = self.opponents[opponent]
randomAction = 0
samples = 1000
for i in range(samples):
#scale chosen by inspection
randomAction +=int(abs(logistic(currOpp.AF*maxRaise,1)))
randomAction /= samples
expectedActions[currOpp.name] = Player.closestInt(enumRaises,randomAction)
for oppAction in expectedActions:
if expectedActions[oppAction]==callSize-1:
continue
else:
impliedPot+=expectedActions[oppAction]
self.impliedPotOdds = callSize/impliedPot
else:
pass
except:
#no CALL or RAISE in histories
pass
# # * Python Numpy Logisitic Distribution #%% import numpy as np from numpy import random import matplotlib.pyplot as pt import seaborn as sb x = random.logistic(loc=1, scale=2, size=(4, 7)) print(x) # %% sb.distplot(random.logistic(size=1000)) pt.show() # %% # * Normal and Logistic distribution sb.distplot(random.normal(scale=2, size=1000), label='normal') sb.distplot(random.logistic(size=1000), label='logistic') pt.show() # %%
def rvs(self, size=None):
return random.logistic(loc=self.loc,
scale=self.scale,
size=self.shape(size))
def logistic_plus(mu, s):
result = logistic(mu, s)
while result < 0:
result = logistic(mu, s)
return result
# logistic distribution is used to describe growth # used extensively in machine learning in logistic regression, neural networks, etc # loc - mean, where the peak is (default 0) # scale - standard deviation, the flatness of distribution (default 1) # size - the shape of the returned array from numpy import random x = random.logistic(loc = 1, scale = 2, size = (2, 3)) print(x) from numpy import random import matplotlib.pyplot as plt import seaborn as sns sns.distplot(random.logistic(size = 1000), hist=False) plt.show()
timecost.append([mid_time-start_time,time.time()-mid_time])
#laplace
start_time=time.time()
a=dsg.laplace(2,2,times)
mid_time=time.time()
b=nr.laplace(2,2,times)
timecost.append([mid_time-start_time,time.time()-mid_time])
#logistic
start_time=time.time()
a=dsg.logistic(2,2,times)
mid_time=time.time()
b=nr.logistic(2,2,times)
timecost.append([mid_time-start_time,time.time()-mid_time])
#lognormal
start_time=time.time()
a=dsg.lognormal(1,1,times)
mid_time=time.time()
b=nr.lognormal(1,1,times)
timecost.append([mid_time-start_time,time.time()-mid_time])
#normal
start_time=time.time()
a=dsg.normal(1,1,times)
mid_time=time.time()
def logistic(size, params):
try:
return random.logistic(params['loc'], params['scale'], size)
except ValueError as e:
exit(e)
#discrete distribution, tells how how manz times something # _can happen given that we know it happens certain amout of time = lam x = random.poisson(lam = 4, size = 100000) sns.distplot(x) # again poisson is similar to normal sns.distplot(random.normal(loc=30, scale=5, size=10000), hist = False, label='normal') sns.distplot(random.poisson(lam=30, size=10000), hist = False, label='poisson') #%% Uniform distribution x = random.uniform(1, 10, size = (1000)) sns.distplot(x, kde = False) #%% logistic and lognormal random.logistic(loc = 1, scale =2, size = 1000) sns.distplot(random.normal(loc = 0, scale =1, size = 1000), hist = False, label = 'Normal') sns.distplot(random.logistic(loc = 0, scale =1, size = 1000), hist= False, label = 'logistic') sns.distplot(random.lognormal(mean = 0.0, sigma = 1, size = 1000)) # can not be negative # ligistic distribution has more area in the tails, atherwise the distributions are similar #%% exponential dist sns.distplot(random.exponential(scale = 2, size = (10, 10))) #%% Vectorization - ufuncs a = [1, 2, 3, 4] b = [10, 20, 30, 40] z = []
import matplotlib.pyplot as plt
import numpy as np
import seaborn as sb
from numpy import random
if __name__ == '__main__':
# arr = np.arange(1, 101)
# sb.distplot(arr)
# print(arr)
arr = random.normal(loc=50, scale=20, size=100)
print(arr)
# sb.histplot(arr, color="lightblue")
sb.distplot(arr, kde=True)
# x = random.binomial(n=100, p=0.8, size=100)
# print(x)
# sb.distplot(x, hist=True, kde=False)
# x = random.poisson(lam=5, size=100)
# print(x)
# sb.distplot(x, hist=True, kde=False)
x = random.logistic(loc=50, scale=10, size=100)
print(x)
# sb.histplot(x, color="darkorange")
sb.distplot(x, kde=True)
plt.show()
def _calculate_noise_for_node(self, node):
loc = 0
scale = power(pi, 2) / 3 * power(self.NOISE, 2)
noise = random.logistic(loc, scale)
return noise
from numpy import random x = random.logistic(loc=1, scale=2, size=(2, 3)) print(x)
for i in range(10):
points = data[data[:, -1] == i]
plt.scatter(points[:, 0], points[:, 1])
plt.show()
np.save("data/2.npy", data)
# 3
data = rdm.uniform([9, 10], [9, 50], (600, 2))
lm = rdm.uniform(0, 2 * np.pi, 600)
data[:, 0] = data[:, 0] + lm * 5
data[:, -1] = data[:, -1] + np.sin(lm) * 50
data = np.vstack((data, rdm.uniform([27, 50], [27, 80], (600, 2))))
lm = np.linspace(0, 2 * np.pi, 600)
data[600:, 0] = data[600:, 0] + lm * 5
data[600:, -1] = data[600:, -1] + np.sin(lm) * 50
data = np.vstack((data, rdm.logistic(34, 0.5, (300, 2))))
data[1200:, -1] = data[1200:, -1] * 4 - 80
data = np.vstack((data, rdm.uniform([15, -35], [20, 30], (300, 2))))
data = np.vstack((data, rdm.normal((50, 90), (2, 8), (300, 2))))
data[:, -1] = data[:, -1] / 3 + 20
tags = np.vstack((np.ones((600, 1)), np.ones(
(600, 1)) * 2, np.ones((300, 1)) * 3, np.ones(
(300, 1)) * 4, np.ones((300, 1)) * 5))
index = np.arange(2100)
data = np.hstack((data, tags))
rdm.shuffle(data)
for i in range(10):
points = data[data[:, -1] == i]
plt.scatter(points[:, 0], points[:, 1])
plt.xlim(0, 70)
plt.ylim(0, 70)
from numpy import random import matplotlib.pyplot as plt import seaborn as sns arr=random.logistic(size=(2,4)) print(arr) sns.distplot(arr,hist=False) plt.show()
1. ye growth ko show karne ke liye use hota hai.
2. ye bahot jaruri hai because aage chal ke ye ml aur neural networking me use hoga.
3. ye continuous probability distribution hai
4. isme 2 parameters hote hain pdf (probability density function) and cdf (cumulative density function)
Logistic distribution arg:
loc: mean , deafult me ye 0 (zero)
scale: standard deviation, default value is 1
size: matrix
'''
from numpy import random as r
import seaborn as sns
import matplotlib.pyplot as plt
logistcVar = r.logistic(loc=1, scale=2, size=(2,3))
print('\n\n',logistcVar,'\n')
# use seaborn to get plto point
# sns.distplot(r.logistic(loc=1, scale=1, size=(100)), hist=False)
# use plt.show() to show graph
sns.distplot(r.normal(loc=1, scale=2, size=(100)), hist=False, label='noraml')
sns.distplot(r.logistic(loc=1, scale=1, size=(100)), hist=False, label='logstic')
plt.show()
'''
Diff b/w Logistc and Normal Distribution
1. logict jo hai ye normal ke comapristion me thoda log distance tak result deta hai.
2. logistic and noraml both center pont result will be like same.
#Generate a random integer from 0 to 100 print(rd.randint(100)) #Generate a 1-D array containing 5 random integers from 0 to 100: print(rd.randint(100, size=(5))) #Generate a 2-D array with 3 rows, each row containing 5 random integers from 0 to 100: print(rd.randint(100, size=(3, 5))) #Return one of the values in an array: print(rd.choice([3, 5, 7, 9])) #Generate a 2-D array that consists of the values in the array parameter (3, 5, 7, and 9): print(rd.choice([3, 5, 7, 9], size=(3, 5))) print(rd.choice([3, 5, 7, 9], p=[0.1, 0.3, 0.6, 0.0], size=(100))) print(rd.choice([3, 5, 7, 9], p=[0.1, 0.3, 0.6, 0.0], size=(3, 5))) # Draw three numbers greater than or equal to 1.0 and less than 2.0 print(rd.uniform(1.0, 2.0, 3)) # Draw three numbers from a normal distribution with mean 0.0 # and standard deviation of 1.0 print(rd.normal(0.0, 1.0, 3)) # Draw three numbers from a logistic distribution with mean 0.0 and scale of 1.0 print(rd.logistic(0.0, 1.0, 3)) # Set seed rd.seed(0) # Generate three random floats between 0.0 and 1.0 print(rd.random(3))
def logistic(self, mean, scale):
'''
Parameters:\n
mean: float. scale: float, >=0.
'''
return r.logistic(mean, scale, self.size)
# logistic distribution is used to describe growth. # used extensively in machine learning in logistic regression, neural networks etc. # it has three parameters: # loc = mean, where peak is. Default = 0. # scale = standard deviation, flatness of distribution. Default 1. # size = Shape of returned array. from numpy import random import matplotlib.pyplot as plt import seaborn as sns arr1 = random.logistic(loc=10, scale=2, size=10) print(arr1) arr2 = random.normal(loc=10, scale=2, size=10) print(arr2) sns.distplot(arr1, hist=False, label='logistic') sns.distplot(arr2, hist=False, label='normal') plt.show() # both are near identical, but logistic has more area under tails. # i.e. possibility of more occurrence of number that are futher away from mean
