Mercurial > hg > nsaunier > traffic-intelligence

comparison python/utils.py @ 86:f03ec4697a09

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
corrected discrete distribution
author Nicolas Saunier <nicolas.saunier@polymtl.ca>
date Tue, 19 Apr 2011 02:01:09 -0400
parents 7f1e54234f96
children 550556378466
comparison
equal deleted inserted replaced
85:7f1e54234f96 86:f03ec4697a09
29 29
30 30
31 class EmpiricalDiscreteDistribution(EmpiricalDistribution): 31 class EmpiricalDiscreteDistribution(EmpiricalDistribution):
32 '''Class to represent a sample of a distribution for a discrete random variable 32 '''Class to represent a sample of a distribution for a discrete random variable
33 ''' 33 '''
34 from numpy.core.fromnumeric import sum
35
34 def __init__(self, categories, counts): 36 def __init__(self, categories, counts):
35 self.categories = categories 37 self.categories = categories
36 self.counts = counts 38 self.counts = counts
37 39
38 def mean(self): 40 def mean(self):
39 from numpy.core.fromnumeric import sum
40 result = [float(x*y) for x,y in zip(self.categories, self.counts)] 41 result = [float(x*y) for x,y in zip(self.categories, self.counts)]
41 return sum(result)/self.nSamples() 42 return sum(result)/self.nSamples()
42 43
43 def var(self, mean = None): 44 def var(self, mean = None):
44 from numpy.core.fromnumeric import sum
45 if not mean: 45 if not mean:
46 m = self.mean() 46 m = self.mean()
47 else: 47 else:
48 m = mean 48 m = mean
49 result = 0. 49 result = 0.
50 squares = [float((x-m)*(x-m)*y) for x,y in zip(self.categories, self.counts)] 50 squares = [float((x-m)*(x-m)*y) for x,y in zip(self.categories, self.counts)]
51 return sum(squares)/(self.nSamples()-1) 51 return sum(squares)/(self.nSamples()-1)
52 52
53 def referenceCounts(self, probability):
54 '''probability is a function that returns the probability of the random variable for the category values'''
55 refProba = [probability(c) for c in self.categories]
56 refProba[-1] = 1-sum(refProba[:-1])
57 refCounts = [r*self.nSamples() for r in refProba]
58 return refCounts, refProba
53 59
54 class EmpiricalContinuousDistribution(EmpiricalDistribution): 60 class EmpiricalContinuousDistribution(EmpiricalDistribution):
55 '''Class to represent a sample of a distribution for a continuous random variable 61 '''Class to represent a sample of a distribution for a continuous random variable
56 with the number of observations for each interval 62 with the number of observations for each interval
57 intervals (categories variable) are defined by their left limits, the last one being the right limit 63 intervals (categories variable) are defined by their left limits, the last one being the right limit