Mercurial > hg > nsaunier > traffic-intelligence

comparison python/utils.py @ 672:5473b7460375

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
moved and rationalized imports in modules
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Tue, 26 May 2015 13:53:40 +0200
parents 849f5f8bf4b9
children 01b89182891a
comparison
equal deleted inserted replaced
671:849f5f8bf4b9 672:5473b7460375
3 3
4 import matplotlib.pyplot as plt 4 import matplotlib.pyplot as plt
5 from datetime import time, datetime 5 from datetime import time, datetime
6 from math import sqrt, ceil, floor 6 from math import sqrt, ceil, floor
7 from scipy.stats import kruskal, shapiro 7 from scipy.stats import kruskal, shapiro
8 from numpy import zeros, array, exp, sum as npsum, arange, cumsum
8 9
9 datetimeFormat = "%Y-%m-%d %H:%M:%S" 10 datetimeFormat = "%Y-%m-%d %H:%M:%S"
10 11
11 ######################### 12 #########################
12 # Enumerations 13 # Enumerations
58 def nSamples(self): 59 def nSamples(self):
59 return sum(self.counts) 60 return sum(self.counts)
60 61
61 def cumulativeDensityFunction(sample, normalized = False): 62 def cumulativeDensityFunction(sample, normalized = False):
62 '''Returns the cumulative density function of the sample of a random variable''' 63 '''Returns the cumulative density function of the sample of a random variable'''
63 from numpy import arange, cumsum
64 xaxis = sorted(sample) 64 xaxis = sorted(sample)
65 counts = arange(1,len(sample)+1) # dtype = float 65 counts = arange(1,len(sample)+1) # dtype = float
66 if normalized: 66 if normalized:
67 counts /= float(len(sample)) 67 counts /= float(len(sample))
68 return xaxis, counts 68 return xaxis, counts
69 69
70 class EmpiricalDiscreteDistribution(EmpiricalDistribution): 70 class EmpiricalDiscreteDistribution(EmpiricalDistribution):
71 '''Class to represent a sample of a distribution for a discrete random variable 71 '''Class to represent a sample of a distribution for a discrete random variable
72 ''' 72 '''
73 from numpy.core.fromnumeric import sum
74
75 def __init__(self, categories, counts): 73 def __init__(self, categories, counts):
76 self.categories = categories 74 self.categories = categories
77 self.counts = counts 75 self.counts = counts
78 76
79 def mean(self): 77 def mean(self):
80 result = [float(x*y) for x,y in zip(self.categories, self.counts)] 78 result = [float(x*y) for x,y in zip(self.categories, self.counts)]
81 return sum(result)/self.nSamples() 79 return npsum(result)/self.nSamples()
82 80
83 def var(self, mean = None): 81 def var(self, mean = None):
84 if not mean: 82 if not mean:
85 m = self.mean() 83 m = self.mean()
86 else: 84 else:
87 m = mean 85 m = mean
88 result = 0. 86 result = 0.
89 squares = [float((x-m)*(x-m)*y) for x,y in zip(self.categories, self.counts)] 87 squares = [float((x-m)*(x-m)*y) for x,y in zip(self.categories, self.counts)]
90 return sum(squares)/(self.nSamples()-1) 88 return npsum(squares)/(self.nSamples()-1)
91 89
92 def referenceCounts(self, probability): 90 def referenceCounts(self, probability):
93 '''probability is a function that returns the probability of the random variable for the category values''' 91 '''probability is a function that returns the probability of the random variable for the category values'''
94 refProba = [probability(c) for c in self.categories] 92 refProba = [probability(c) for c in self.categories]
95 refProba[-1] = 1-sum(refProba[:-1]) 93 refProba[-1] = 1-npsum(refProba[:-1])
96 refCounts = [r*self.nSamples() for r in refProba] 94 refCounts = [r*self.nSamples() for r in refProba]
97 return refCounts, refProba 95 return refCounts, refProba
98 96
99 class EmpiricalContinuousDistribution(EmpiricalDistribution): 97 class EmpiricalContinuousDistribution(EmpiricalDistribution):
100 '''Class to represent a sample of a distribution for a continuous random variable 98 '''Class to represent a sample of a distribution for a continuous random variable
166 # return smoothed 164 # return smoothed
167 165
168 def kernelSmoothing(x, X, Y, weightFunc, halfwidth): 166 def kernelSmoothing(x, X, Y, weightFunc, halfwidth):
169 '''Returns the smoothed estimate of (X,Y) at x 167 '''Returns the smoothed estimate of (X,Y) at x
170 Sum_x weight(sample_x,x) * y(x)''' 168 Sum_x weight(sample_x,x) * y(x)'''
171 from numpy import zeros, array
172 weights = array([weightFunc(x,observedx, halfwidth) for observedx in X]) 169 weights = array([weightFunc(x,observedx, halfwidth) for observedx in X])
173 if sum(weights)>0: 170 if sum(weights)>0:
174 return sum(weights*Y)/sum(weights) 171 return sum(weights*Y)/sum(weights)
175 else: 172 else:
176 return 0 173 return 0
180 return 1. 177 return 1.
181 else: 178 else:
182 return 0. 179 return 0.
183 180
184 def gaussian(center, x, halfwidth): 181 def gaussian(center, x, halfwidth):
185 from numpy import exp
186 return exp(-((center-x)/halfwidth)**2/2) 182 return exp(-((center-x)/halfwidth)**2/2)
187 183
188 def epanechnikov(center, x, halfwidth): 184 def epanechnikov(center, x, halfwidth):
189 diff = abs(center-x) 185 diff = abs(center-x)
190 if diff<halfwidth: 186 if diff<halfwidth: