sergeyf

import numpy as np
import matplotlib.pyplot
mu, sigma = 3., 1. # mean and standard deviation
s = np.random.lognormal(mu, sigma, 10000)
log_s = np.log(s)

subplot(211)
count,bins,_ = hist(s, 100, normed=True, align='mid')
x = np.linspace(min(bins), max(bins), 10000)
pdf = (np.exp(-(np.log(x) - mu)**2 / (2 * sigma**2)) / (x * sigma * np.sqrt(2 * np.pi)))

sergeyf

from scipy.stats import rv_continuous
from scipy.special import gammaln, gammaincinv, gammainc
from numpy import log,exp

class igamma_gen(rv_continuous):
  def _pdf(self, x, a, b):
      return exp(self._logpdf(x,a,b))
  def _logpdf(self, x, a, b):
      return a*log(b) - gammaln(a) -(a+1)*log(x) - b/x
  def _cdf(self, x, a, b):

sergeyf

import numpy as np
from sklearn.feature_extraction import image
from sklearn.cluster import MiniBatchKMeans
from sklearn import cross_validation, svm, datasets
from sklearn.datasets import fetch_olivetti_faces, fetch_mldata
from matplotlib import pylab as pl

def HIK_kernel(X,Y):
    return np.array([[np.sum(np.minimum(x,y)) for y in Y] for x in X]) 
    

sergeyf

'''
Based on: https://gist.github.com/bshillingford/6259986edca707ca58dd
Modified to work on Windows by: Sergey Feldman
Jan 17, 2016

Requirements: pdflatex, bibtex
'''

import requests
import lxml.html as html

sergeyf

import numpy as np
from scipy.spatial.distance import pdist, squareform

# function that converts categorical variable 
# into a one-hot encoding
def one_hot_encoding(x):
    n = len(x)    
    min_category = np.min(x)
    max_category = np.max(x)
    num_categories = max_category - min_category + 1

sergeyf

import numpy as np

def f_of_x(X,w):
    n,d = X.shape
    X_dot_w = np.dot(X,w)
    y = np.zeros(n)
    # the inner product goes through a sin
    # or a cos, depending on simple condition
    cos_flag = X[:,0] < 0.0
    sin_flag = ~cos_flag

sergeyf

import numpy as np
from matplotlib import pylab as plt
#from mpltools import style # uncomment for prettier plots
#style.use(['ggplot'])

# generate all bernoulli rewards ahead of time
def generate_bernoulli_bandit_data(num_samples,K):
    CTRs_that_generated_data = np.tile(np.random.rand(K),(num_samples,1))
    true_rewards = np.random.rand(num_samples,K) < CTRs_that_generated_data
    return true_rewards,CTRs_that_generated_data

sergeyf

import numpy as np

# the function
def f_of_x(X, w):
    n,d = X.shape
    X_dot_w = np.dot(X,w)
    y = np.zeros(n)
    # the inner product randomly goes through a sin
    # or a cos
    cos_flag = np.random.randn(n) < 0.0

sergeyf

'''
References: 
    https://medium.com/teconomics-blog/using-ml-to-resolve-experiments-faster-bd8053ff602e
    https://insightr.wordpress.com/2017/06/28/cross-fitting-double-machine-learning-estimator/
    https://arxiv.org/pdf/1608.00060.pdf
'''

import numpy as np

from sklearn.linear_model import LassoCV, LinearRegression, BayesianRidge, LogisticRegression

sergeyf

import numpy as np

class Data_generator(object):
    
    def __init__(self,K,d,reward_type='binary'):
        
        self.d = d # dimension of the feature vector
        self.K = K # number of bandits
        self.reward_type = reward_type