def Var3(t, mu=None):
if mu is None:
mu = thinkstats.Mean(t)
# compute the squared deviations and return their mean.
dev2 = [(x - mu)**3 for x in t]
var = thinkstats.Mean(dev2)
return var
def Sample():
exp = randvar.Exponential(p_lam)
sample = sorted([exp.generate() for x in range(6)])
mean = ts.Mean(sample)
median = ts.Mean(sample[2:4])
return 1.0/mean, math.log(2)/median
def Sample():
normal = randvar.Normal(0, 1)
sample = [normal.generate() for x in range(6)]
mean = ts.Mean(sample)
dev2 = [(x - mean)**2 for x in sample]
sn = sum(dev2) / float(len(dev2))
sn_1 = sum(dev2) / float(len(dev2) - 1)
return sn, sn_1
def runtest(name, actual1, actual2):
print name
# observed delta
mu1, mu2 = thinkstats.Mean(actual1), thinkstats.Mean(actual2)
delta = abs(mu1 - mu2)
n, m = len(actual1), len(actual2)
model = actual1 + actual2
cdf, pvalue = PValue(model, model, n, m, delta)
print 'n:', n
print 'm:', m
print 'mu1', mu1
print 'mu2', mu2
print 'delta', delta
print 'p-value', pvalue
PlotCdf(name, cdf, delta)
def Resample(t1, t2, n, m):
sample1 = [random.choice(t1) for i in range(n)]
sample2 = [random.choice(t2) for i in range(m)]
delta = thinkstats.Mean(sample1) - thinkstats.Mean(sample2)
return delta
def Sample():
normal = randvar.Normal(0, 1)
sample = sorted([normal.generate() for x in range(6)])
return ts.Mean(sample), ts.Mean(sample[2:4])
import thinkstats.thinkstats as ts import math pumpkins = [200, 250, 500, 550, 2000, 2500] print 'Mean of pumpkins:', ts.Mean(pumpkins), 'g' print 'Variance of pumpkins:', ts.Var(pumpkins), 'g^2' print 'Standard Dev. of pumpkins:', math.sqrt(ts.Var(pumpkins)), 'g'
def PearsonSkewness(t):
return 3.0 * (thinkstats.Mean(t) - Median(t)) / math.sqrt(
thinkstats.Var(t))