def cdf(self, val):
if isinstance(val, (float, int)):
if val < self.minimum:
_cdf = 0.
else:
_cdf = gammainc(self.k, val / self.theta)
else:
_cdf = np.zeros(val.size)
_cdf[val >= self.minimum] = gammainc(self.k, val[val >= self.minimum] / self.theta)
return _cdf
def chisquare_solve(XGUESS, P, V):
PGUESS = gammainc(V / 2, V * XGUESS / 2) # incomplete Gamma function
PDIFF = np.abs(PGUESS - P) # error in calculated P
TOL = 1E-4
if PGUESS >= 1 - TOL: # if P is very close to 1 (i.e. a bad guess)
PDIFF = XGUESS # then just assign some big number like XGUESS
return PDIFF
def chisquare_solve(XGUESS,P,V): PGUESS = gammainc(V/2, V*XGUESS/2) # incomplete Gamma function PDIFF = np.abs(PGUESS - P) # error in calculated P TOL = 1E-4 if PGUESS >= 1-TOL: # if P is very close to 1 (i.e. a bad guess) PDIFF = XGUESS # then just assign some big number like XGUESS return PDIFF
def chisquare_solve(xguess, p, v):
"""
Chisqure_solve
Return the difference between calculated percentile and P.
Written January 1998 by C. Torrence
"""
pguess = gammainc(v / 2, v * xguess / 2)
pdiff = np.abs(pguess - p)
if pguess >= 1 - 1e-4:
pdiff = xguess
return pdiff
def chisquare_solve(xguess,p,v):
"""
Chisqure_solve
Return the difference between calculated percentile and P.
Written January 1998 by C. Torrence
"""
pguess = gammainc(v/2,v*xguess/2)
pdiff = np.abs(pguess - p)
if pguess >= 1-1e-4:
pdiff = xguess
return pdiff
def AbsorptionConditionalProb(a, b):
''' probability that F_ti+1 is absorbed by the 0 barrier conditional on inital value S0 '''
cprob = 1. - gammainc(b / 2, a / 2) # formula (2.10), scipy gammainc is already normalized by gamma(b)
return cprob
def theoretical_distribution(self):
x = np.sort(self.sample)
y = np.array([gammainc(self.freedom_degrees / 2, i / 2) for i in x])
plt.plot(x, y, 'r-')