def getActuallyKnownRatio(self, ct=None):
if ct==None:
ct=time()
tot=sum(1 for v,t in self.actuallyKnownHist if t>ct-actuallyKnownTime)
actuallyKnown=sum(1 for v,t in self.actuallyKnownHist if t>ct-actuallyKnownTime and v)
if tot==0:
return 0., 0., 1.
else:
r=float(actuallyKnown)/tot
alpha=1-conf
return (r,0. if actuallyKnown==0 else btdtri(actuallyKnown,tot-actuallyKnown+1,alpha/2),
1. if actuallyKnown==tot else btdtri(actuallyKnown+1,tot-actuallyKnown,1-alpha/2))
def ppf(self, x):
"""
Computes the percent point function of the distribution at the point(s)
x. It is defined as the inverse of the CDF. y = ppf(x) can be
interpreted as the argument y for which the value of the cdf(x) is equal
to y. Essentially that means the random varable y is the place on the
distribution the CDF evaluates to x.
Parameters
----------
x: array, dtype=float, shape=(m x n), bounds=(0,1)
The value(s) at which the user would like the ppf evaluated.
If an array is passed in, the ppf is evaluated at every point
in the array and an array of the same size is returned.
Returns
-------
ppf: array, dtype=float, shape=(m x n)
The ppf at each point in x.
"""
if (x <= 0).any() or (x >= 1).any():
raise ValueError('all values in x must be between 0 and 1, \
exclusive')
ppf = btdtri(self.alpha, self.beta, x)
return ppf
def probf_baharev(df1, df2, noncen, fcrit):
x = 1 - special.btdtri(df1, df2, fcrit)
eps = 1.0e-7
itr_cnt = 0
f = None
while itr_cnt <= 10:
mu = noncen / 2.0
ql = poisson.ppf(eps, mu)
qu = poisson.ppf(1 - eps, mu)
k = qu
c = beta.cdf(x, df1 + k, df2)
d = x * (1.0 - x) / (df1 + k - 1.0) * beta.pdf(x, df1 + k - 1, df2, 0)
p = poisson.pmf(k, mu)
f = p * c
p = k / mu * p
k = qu - 1
while k >= ql:
c = c + d
d = (df1 + k) / (x * (df1 + k + df2 - 1)) * d
f = f + p * c
p = k / mu * p
k = k - 1
itr_cnt = itr_cnt + 1
if (itr_cnt == 11):
print("newton iteration failed")
return f
def ppf(self, x):
"""
Computes the percent point function of the distribution at the point(s)
x. It is defined as the inverse of the CDF. y = ppf(x) can be
interpreted as the argument y for which the value of the cdf(x) is equal
to y. Essentially that means the random varable y is the place on the
distribution the CDF evaluates to x.
Parameters
----------
x: array, dtype=float, shape=(m x n), bounds=(0,1)
The value(s) at which the user would like the ppf evaluated.
If an array is passed in, the ppf is evaluated at every point
in the array and an array of the same size is returned.
Returns
-------
ppf: array, dtype=float, shape=(m x n)
The ppf at each point in x.
"""
if (x <=0).any() or (x >=1).any():
raise ValueError('all values in x must be between 0 and 1, \
exclusive')
ppf = btdtri(self.alpha, self.beta, x)
return ppf
def main():
parcellations = putils.get_cammoun_schaefer(data_dir=ROIDIR)
# output dataframe
cols = ['parcellation', 'scale', 'spintype', 'n_sig']
data = pd.DataFrame(columns=cols)
# let's run all our parcellations
for parcellation, annotations in parcellations.items():
print(f'PARCELLATION: {parcellation}')
for scale in annotations:
for spintype in SPINTYPES:
data = data.append(run_null(parcellation, scale, spintype),
ignore_index=True)
# now calculate parametric null
nsdata = load_data(parcellation, scale, spintype)
corrs = np.corrcoef(nsdata.T)
# this calculates the correlation value for p < ALPHA cutoff
ab = (len(nsdata) / 2) - 1
cutoff = 1 - (special.btdtri(ab, ab, ALPHA / 2) * 2)
# now add the parametric null to our giant summary dataframe
data = data.append(pd.DataFrame({
'parcellation': parcellation,
'scale': scale,
'spintype': 'naive-para',
'n_sig': np.sum(np.triu(corrs > cutoff, k=1))
}, index=[0]), ignore_index=True)
data.to_csv(NSDIR / 'ns_summary.csv.gz', index=False)
def _ppf(self, rho, a, b, p, q):
# subclass the _ppf method (returns the inverse cdf of the
# beinf distribution). NOTE: This is not a true inverse
# when p!=0 since the same SIC value can be returned
# for a range of probabilities at the endpoints
if np.all(p==1):
gamma_ = 0.0
else:
gamma_ = (rho - p*(1-q))/(1-p)
condlist=[np.logical_and(rho>=0., rho<=p*(1-q)),
np.logical_and(rho>p*(1-q),rho<(1-p*q)),
np.logical_and(rho>=1-p*q, rho<=1.)]
choicelist=[0.0, special.btdtri(a,b,gamma_), 1.0]
return np.select(condlist, choicelist)
def _ppf(self, q, a, b):
return special.btdtri(a, b, q)
def efficiency_err(group, threshold, PUWP, ISOWP, upper=False):
tot = group.shape[0]
if PUWP == '99':
if ISOWP == '10':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP10 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
elif ISOWP == '15':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP15 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
elif ISOWP == '20':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP20 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
elif ISOWP == '90':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP90 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
elif ISOWP == '95':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP95 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
else:
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP99 == True) &
(group.cl3d_pubdt_passWP99 == True)].shape[0]
elif PUWP == '95':
if ISOWP == '10':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP10 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
elif ISOWP == '15':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP15 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
elif ISOWP == '20':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP20 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
elif ISOWP == '90':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP90 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
elif ISOWP == '95':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP95 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
else:
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP99 == True) &
(group.cl3d_pubdt_passWP95 == True)].shape[0]
else:
if ISOWP == '10':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP10 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
elif ISOWP == '15':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP15 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
elif ISOWP == '20':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP20 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
elif ISOWP == '90':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP90 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
elif ISOWP == '95':
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP95 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
else:
sel = group[(group.cl3d_pt_c3 > threshold)
& (group.cl3d_isobdt_passWP99 == True) &
(group.cl3d_pubdt_passWP90 == True)].shape[0]
# clopper pearson errors --> ppf gives the boundary of the cinfidence interval, therefore for plotting we have to subtract the value of the central value float(sel)/float(tot)!!
alpha = (1 - 0.9) / 2
if upper:
if sel == tot:
return 0.
else:
return abs(
btdtri(sel + 1, tot - sel, 1 - alpha) -
float(sel) / float(tot))
else:
if sel == 0:
return 0.
else:
return abs(
float(sel) / float(tot) - btdtri(sel, tot - sel + 1, alpha))
def _ppf(self, qloc, a, b):
return special.btdtri(a, b, qloc)
def quantile(self, p):
return btdtri(
self.N[1], self.N[0],
p) # Bug: do not call btdtri with (0.5,0.5,0.5) in scipy < 0.9
def quantile(self, p):
return btdtri(self.N[1], self.N[0], p) # Bug: do not call btdtri with (0.5,0.5,0.5) in scipy < 0.9
def _ppf(self, q):
a, b = np.loadtxt(os.path.join(FILE_DIR, 'distr_par.txt'), delimiter = ',')
return special.btdtri(a, b, q)
def quantile(self, p):
return btdtri(self.N[1], self.N[0], p)
def quantile(self, *q):
# check array for numpy structure
q = check_array(q, reduce_args=True, ensure_1d=True)
return sc.btdtri(self.alpha, self.beta, q)
def quantile(self, p):
return btdtri(self.params[1], self.params[0], p)
def btdtri_comp(a, b, p):
return btdtri(a, b, 1-p)
def btdtri_comp(a, b, p):
return btdtri(a, b, 1-p)
def quantile(self, p):
"""Return the p quantile of the Beta posterior (using :func:`scipy.stats.btdtri`).
- Used only by :class:`BayesUCB` and :class:`AdBandits` so far.
"""
return btdtri(self.N[1], self.N[0], p)
