def rapid(self):
a, b = 5, 100
x_sim = np.linspace(wald.ppf(0.01), wald.ppf(0.99), self.n_intervals)
y = np.exp(wald.pdf(x_sim) / 2)
y[2:] = y[:-2]
y[:2] = 1
return y
def rapid(self):
a, b = 5, 100
y = np.linspace(wald.ppf(0.01), wald.ppf(0.99),
self.n_effective_intervals)
y = np.exp(wald.pdf(y) / 1.6)
y[2:] = y[:-2]
y[:2] = 1
effect = np.ones(self.n_intervals)
effect[0:self.n_effective_intervals] = y
return effect
'/home/fearofchou/data/KKBOX/libfm/tr_te_data_BS/split_by_month/short_term_playcount/ID_02*'
)
with open(
'/home/fearofchou/data/KKBOX/libfm/tr_te_data_BS/split_by_month/short_term_playcount/rel_target_02.test'
) as f:
true = np.array(f.readlines()).astype('int')
with open(
'/home/fearofchou/data/KKBOX/fearure_extraction/song_age/rel_songage_201212201301'
) as f:
sa = np.array(f.readlines()).astype('int')
f_sa = sa
for idx, val in enumerate(sa):
if val > 12:
f_sa[idx] = 12
if val < 0:
f_sa[idx] = 0
pred = {}
for fn in fl:
with open(fn) as f:
pred[fn.split('/')[-1]] = np.array(f.readlines()).astype('float')
x = pl.frange(0, 1.3, 0.1)
wald_dis = wald.pdf(x, 0, 0.5)
f_sa = f_sa + 1
for fn in pred.keys():
wald_pred = (wald_dis[f_sa] * pred[fn] + pred[fn]) / 2
print fn
print np.sqrt(((true - wald_pred)**2).mean())
def density(self, x, _mu=None, _sigma=None):
return wald.pdf(x, loc=_mu, scale=_sigma)
def z(a):
return wald.pdf(1-a)
from scipy.stats import wald import matplotlib.pyplot as plt fig, ax = plt.subplots(1, 1) # Calculate a few first moments: mean, var, skew, kurt = wald.stats(moments='mvsk') # Display the probability density function (``pdf``): x = np.linspace(wald.ppf(0.01), wald.ppf(0.99), 100) ax.plot(x, wald.pdf(x), 'r-', lw=5, alpha=0.6, label='wald pdf') # Alternatively, the distribution object can be called (as a function) # to fix the shape, location and scale parameters. This returns a "frozen" # RV object holding the given parameters fixed. # Freeze the distribution and display the frozen ``pdf``: rv = wald() ax.plot(x, rv.pdf(x), 'k-', lw=2, label='frozen pdf') # Check accuracy of ``cdf`` and ``ppf``: vals = wald.ppf([0.001, 0.5, 0.999]) np.allclose([0.001, 0.5, 0.999], wald.cdf(vals)) # True # Generate random numbers: r = wald.rvs(size=1000)
'''
import glob
fl = glob.glob('/home/fearofchou/data/KKBOX/libfm/tr_te_data_BS/split_by_month/short_term_playcount/ID_02*')
with open('/home/fearofchou/data/KKBOX/libfm/tr_te_data_BS/split_by_month/short_term_playcount/rel_target_02.test') as f:
true = np.array(f.readlines()).astype('int')
with open('/home/fearofchou/data/KKBOX/fearure_extraction/song_age/rel_songage_201212201301') as f:
sa = np.array(f.readlines()).astype('int')
f_sa = sa
for idx,val in enumerate(sa):
if val>12:
f_sa[idx]=12
if val<0:
f_sa[idx]=0
pred = {}
for fn in fl:
with open(fn) as f:
pred[fn.split('/')[-1]] = np.array(f.readlines()).astype('float')
x = pl.frange(0,1.3,0.1)
wald_dis = wald.pdf(x,0,0.5)
f_sa =f_sa+1
for fn in pred.keys():
wald_pred=(wald_dis[f_sa]*pred[fn] + pred[fn])/2
print fn
print np.sqrt( ((true-wald_pred)**2).mean() )
def data_prob(pm, y):
return wald.pdf(y, pm['mu'], pm['lambda'])
from scipy.stats import wald
import matplotlib.pyplot as plt
fig, ax = plt.subplots(1, 1)
# Calculate a few first moments:
mean, var, skew, kurt = wald.stats(moments='mvsk')
# Display the probability density function (``pdf``):
x = np.linspace(wald.ppf(0.01),
wald.ppf(0.99), 100)
ax.plot(x, wald.pdf(x),
'r-', lw=5, alpha=0.6, label='wald pdf')
# Alternatively, the distribution object can be called (as a function)
# to fix the shape, location and scale parameters. This returns a "frozen"
# RV object holding the given parameters fixed.
# Freeze the distribution and display the frozen ``pdf``:
rv = wald()
ax.plot(x, rv.pdf(x), 'k-', lw=2, label='frozen pdf')
# Check accuracy of ``cdf`` and ``ppf``:
vals = wald.ppf([0.001, 0.5, 0.999])
np.allclose([0.001, 0.5, 0.999], wald.cdf(vals))
# True
# Generate random numbers: