def Ftest_pvalue(d1, d2):
df1 = len(d1) - 1
df2 = len(d2) - 1
F = stats.variance(d1) / stats.variance(d2)
single_tailed_pval = ss.f.cdf(F, df1, df2)
double_tailed_pval = single_tailed_pval * 2
return (single_tailed_pval, double_tailed_pval)
def Ftest_pvalue(d1,d2):
df1 = len(d1) - 1
df2 = len(d2) - 1
F = stats.variance(d1) /stats.variance(d2)
single_tailed_pval = ss.f.cdf(F,df1,df2)
double_tailed_pval = single_tailed_pval * 2
return double_tailed_pval
print("F-test value:",Ftest_pvalue( d1,d2))
def krige_old(xpt, ypt, zpt, rg, x, y, vtype='spher'):
""" krige function to interpolate over a vector of points of x,y coords
using the base points xpt ypt and the vario distance rg (range)"""
#print 'geom kr l 522',len(xpt),rg
n = len(x)
z0 = zeros(n)
vari = 0.5 * variance(zpt)
#print 'vari',vari
def gamma(vtype, d, rg): # OA 25/10/18
if vtype in ['spher', None]:
gam = 3 / 2. * d / rg - 1 / 2. * (d / rg)**3.
gam[d > rg] = 1
elif vtype == 'gauss':
gam = exp(-(d / rg)**2)
elif vtype == 'gauss3':
gam = exp(-(d / rg)**3)
return gam
for i in range(n):
x0 = x[i]
y0 = y[i]
d = sqrt((x0 - xpt)**2 + (y0 - ypt)**2)
ind = argsort(d)
x1, y1, z1 = xpt[ind[:16]], ypt[ind[:16]], zpt[ind[:16]]
# select closest 16 points
xm1, ym1 = meshgrid(x1, y1)
d = sqrt((xm1 - transpose(xm1))**2. + (ym1 - transpose(ym1))**2.)
gam = gamma(vtype, d, rg) # OA 25/10/18
l1 = len(x1) + 1
A = ones((l1, l1))
#starting the A matrix
A[:-1, :-1] = gam * vari
A[-1, -1] = 0.
d = sqrt((x0 - x1)**2. + (y0 - y1)**2.)
gam = gamma(vtype, d, rg) # OA 25/10/18
B = ones((l1, 1))
B[:-1, 0] = gam * vari
lb = solve(A, B)
z0[i] = sum(z1 * transpose(lb[:-1]))
#z0 = clip(z0,min(zpt)*0.9,max(zpt)*1.1)
return z0
import constante
from funciones import normal_por_aceptacion_rechazo
z = normal_por_aceptacion_rechazo(media=35, de=5)
hist_data = [z]
# ploteo data
fig = ff.create_distplot(hist_data, [""], bin_size=.01, curve_type='normal')
fig['layout'].update(title='Normal empirica vs Normal de python')
py.plot(fig, filename='normal empirica vs normal de python')
# Mostramos media, varianza y moda muestrales y teoricos
media = st.mean(z)
varianza = st.variance(z)
moda = max(set(z), key=z.count)
print("Media muestral: {0} Varianza muestral: {1} Moda muestral: {2}".format(media, varianza, moda))
print("Media teorica: {0} Varianza teorica: {1} Moda teorica: {2}".format(35, 5*5, 35))
# RESPUESTA 5
from funciones import gcl_uniforme
import constante
# genero numero aleatorios uniformes
x_n = constante.SEMILLA
uniformes = []
empiricos = []
for _ in range(constante.CANT_EXPERIMENTOS):
x_n = gcl_uniforme(x_n)
def featurescalculator(sigbufs, n):
Desface = 12000
x = 0 + Desface # desface en # de muestras donde inicia el examen
aumento = 0
segundos = int(
(len(sigbufs[3, :]) - Desface) / 200) # numero de segundos del examen
Features = np.empty(((n - 5) * segundos, 32)) # matriz de caracteristicas
for a in np.arange(segundos):
for i in np.arange((n - 5)): #n-2 numero de canales
warnings.filterwarnings("ignore")
#TIEMPO
minimo = scipy.stats.tmin(sigbufs[i, x:x + 200])
maximo = scipy.stats.tmax(sigbufs[i, x:x + 200])
kurto = scipy.stats.kurtosis(sigbufs[i, x:x + 200])
energ = energia(sigbufs[i, x:x + 200])
sha = shannon(sigbufs[i, x:x + 200])
#DWT
cA5, cD4, cD3, cD2, cD1 = pywt.wavedec(sigbufs[i, x:x + 200],
'db4',
level=4)
varianzaA5 = stats.variance(cA5)
energA5 = energia(cA5)
shaA5 = shannon(cA5)
actiA5 = hjorth(cA5)
varianzaD4 = stats.variance(cD4)
energD4 = energia(cD4)
rD4 = renyi(cD4)
shaD4 = shannon(cD4)
EHD4 = hurst(cD4)
actiA4 = hjorth(cD4)
varianzaD3 = stats.variance(cD3)
desviacionD3 = stats.stdev(cD3)
energD3 = energia(cD3)
rD3 = renyi(cD3)
apenD3 = ApEn(cD3, 2, 3)
shaD3 = shannon(cD3)
minimoD2 = scipy.stats.tmin(cD2)
maximoD2 = scipy.stats.tmax(cD2)
desviacionD2 = stats.stdev(cD2)
kurtoD2 = scipy.stats.kurtosis(cD2)
energD2 = energia(cD2)
rD2 = renyi(cD2)
shaD2 = shannon(cD2)
minimoD1 = scipy.stats.tmin(cD1)
maximoD1 = scipy.stats.tmax(cD1)
rD1 = renyi(cD1)
#FFT
nee = len(sigbufs[i, x:x + 200]) # tamaño
Y = fft(sigbufs[i, x:x + 200]) / nee
Yn = abs(Y)
mediaf = stats.mean(Yn)
#print (signal_labels[i])
Features[i + aumento] = [
minimo, maximo, kurto, energ, sha, varianzaA5, energA5, shaA5,
actiA5, varianzaD4, energD4, rD4, shaD4, EHD4, actiA4,
varianzaD3, desviacionD3, energD3, rD3, apenD3, shaD3,
minimoD2, maximoD2, desviacionD2, kurtoD2, energD2, rD2, shaD2,
minimoD1, maximoD1, rD1, mediaf
]
#Labels=signal_labels[i]
#print (Labels)
x = x + 200
aumento = aumento + 18 ##16 -- n-2, 21 -- n-4, 15 -- n-3, 19 -- n-4, 43 ---- n-8
return Features
