def pValForGi(gVal):
""" This function uses the pdf module to calculate p values.
Parameters:
Name Type Description
gVal Float A Gi statistic value (standard normal variate.
Output:
Name Type Description
pVal Float A value from the normal or t probability distribution.
"""
if gVal >= 0:
if gVal < 6.0:
pVal = 1.0-pdf.zprob(gVal)
else:
pVal = pdf.tpvalue(gVal,3000)
else:
if gVal > -6.0:
pVal = pdf.zprob(gVal)
else:
pVal = pdf.tpvalue(gVal,3000)
return 2.0*pVal
def __init__(self, y, w, permutations=0, varAssumption="normalization"):
self.variable = y
if y.t == 1: # check for cs variable
n = y.n
y = reshape(y, (n, 1))
self.varAssumption = varAssumption
self.iMoments(w)
self.permutations = permutations
self.w = w
yd = y - mean(y)
den = matrixmultiply(transpose(yd), yd)
print sum(sum(den == 0))
# from here down changes for permutations (not done yet)
# check for permutation and then permutate the yd
ylag = lag(yd, w)
self.ylag = lag(y, w)
num = matrixmultiply(transpose(yd), ylag)
if len(shape(num)) > 1:
self.mi = diag(num / den)
self.zi = (self.mi - self.ei) / self.si
n, t = shape(ylag)
else:
n = len(y)
t = 1
self.mi = num / den
self.zi = (self.mi - self.ei) / self.si
self.npvalue = [1 - pdf.zprob(abs(x)) for x in self.zi]
if permutations:
Message(
"Moran's I, %d permutations for %d regions and %d time periods"
% (permutations, n, t))
mip = []
count = zeros([1, t])
for iter in range(permutations):
ylag = lag(permutate(yd), w)
num = matrixmultiply(transpose(yd), ylag)
if len(shape(num)) > 1:
mi = diag(num / den)
else:
mi = num / den
count += self.extreme(mi)
mip.append(mi)
self.ppvalue = count / (permutations * 1.0)
def __init__(self,y,w,permutations=0,varAssumption = "normalization"):
self.variable = y
if y.t == 1: # check for cs variable
n=y.n
y = reshape(y,(n,1))
self.varAssumption = varAssumption
self.iMoments(w)
self.permutations=permutations
self.w = w
yd = y - mean(y)
den = matrixmultiply(transpose(yd),yd)
print sum(sum(den==0))
# from here down changes for permutations (not done yet)
# check for permutation and then permutate the yd
ylag = lag(yd,w)
self.ylag = lag(y,w)
num = matrixmultiply(transpose(yd),ylag)
if len(shape(num)) > 1:
self.mi = diag(num/den)
self.zi = (self.mi - self.ei)/self.si
n,t=shape(ylag)
else:
n=len(y)
t=1
self.mi = num/den
self.zi = (self.mi - self.ei)/self.si
self.npvalue = [ 1 - pdf.zprob(abs(x)) for x in self.zi]
if permutations:
Message("Moran's I, %d permutations for %d regions and %d time periods"
%(permutations,n,t))
mip = []
count =zeros([1,t])
for iter in range(permutations):
ylag=lag(permutate(yd),w)
num = matrixmultiply(transpose(yd),ylag)
if len(shape(num)) > 1:
mi = diag(num/den)
else:
mi = num/den
count += self.extreme(mi)
mip.append(mi)
self.ppvalue = count/(permutations * 1.0)
def __init__(self,y,w,permutations=0,varAssumption = "normalization"):
self.varAssumption = varAssumption
self.w=w
self.variable = y
n=y.n
self.n=n
self.cMoments(w)
self.iMat=ones((n,n))
res=map(self.calc,transpose(y))
self.gc=array(res)
self.zc = (self.gc - 1) / self.sc
self.zpvalue = [ 1 - pdf.zprob(abs(x)) for x in self.zc]
self.permutations = permutations
if permutations:
count = ones([1,y.t])
resperm=[]
iterations=range(permutations)
for it in iterations:
yp=permutate(y)
gc = map(self.calc,transpose(yp))
resperm.append(gc)
count += self.extreme(gc)
self.ppvalue = count/(permutations * 1.0 + 1.0)