def test_1D_density():
np.random.seed(0)
dist = norm(0, 1)
X = dist.rvs((5000, 1))
X2 = np.linspace(-5, 5, 10).reshape((10, 1))
true_dens = dist.pdf(X2[:, 0]) * X.shape[0]
classifiers = [
KNeighborsDensity(method='simple', n_neighbors=250),
KNeighborsDensity(method='bayesian', n_neighbors=250)
]
for clf in classifiers:
yield (check_1D_density, clf, X, X2, true_dens, 100)
def NN_bayesian_density(x, y, NN, grid_size):
"""
Function to compute the density of a distribution of particles
using the K-Nearest Neighboors
method from:
http://www.astroml.org/modules/generated/astroML.density_estimation.KNeighborsDensity.html#astroML.density_estimation.KNeighborsDensity
See Ivezic 10? for the details on how the algorithm works.
Input:
------
x : 1D numpy.array
Array with the x-coordinates of the data.
y : 1D numpy.array
Array with the y-coordinates of the data.
NN : int
Number of neighboors to compute the desnity.
grid_size : int
Grid size in which the density is going to be evaluated.
"""
assert len(x) == len(y), "Input data have different size"
assert type(NN) == int, "NN should be of type int"
assert type(grid_size) == int, "grid_zise should be of type int"
# Grid parameters
Nx = grid_size
Ny = grid_size
xmin, xmax = (min(x), max(x))
ymin, ymax = (min(y), max(y))
# Making a grid
Xgrid = np.vstack(map(np.ravel, np.meshgrid(np.linspace(xmin, xmax, Nx),\
np.linspace(ymin, ymax, Ny)))).T
# Putting data in 2d-array
X = np.array([x, y]).T
# Computing the density
knn = KNeighborsDensity('bayesian', NN)
dens_KNN = knn.fit(X).eval(Xgrid).reshape((Ny, Nx))
return dens_KNN
def __init__(self, evts, t1, t2, k=10000):
self.evts = evts
#Decide the bandwidth
self.t1 = t1
self.t2 = t2
self.dt = self.t2 - self.t1
#Perform the Kernel Density Estimation
knd = KNeighborsDensity('bayesian', n_neighbors=k)
sys.stderr.write("Fitting...")
knd.fit(evts[:, numpy.newaxis])
sys.stderr.write("done")
#Evaluate the KDE and interpolate it
sys.stderr.write("Evaluating...")
x_grid = numpy.arange(self.t1, self.t2, 10.0)
pdf = knd.eval(x_grid[:, numpy.newaxis]) / evts.shape[0]
sys.stderr.write("done")
sys.stderr.write("Interpolating...")
self.model = interpolate.InterpolatedUnivariateSpline(x_grid, pdf, k=1)
sys.stderr.write("done")
setup_text_plots(fontsize=8, usetex=False)
ra, dec = np.loadtxt('simbad3.tsv',usecols=(5,6),unpack=True)
X = np.vstack((dec,ra)).T
k=500
Nx = 300
Ny = 300
xmin, xmax = (-30, 70)
ymin, ymax = (270, 0)
#------------------------------------------------------------
# Evaluate density
Xgrid = np.vstack(map(np.ravel, np.meshgrid(np.linspace(xmin, xmax, Nx),
np.linspace(ymin, ymax, Ny)))).T
knn = KNeighborsDensity('simple', k)
dens = knn.fit(X).eval(Xgrid).reshape((Ny, Nx))
plt.figure(figsize=(9,9))
plt.scatter(X[:, 1], X[:, 0], s=2, lw=0, c='r')
plt.imshow(dens.T, origin='lower',
extent=(ymin, ymax, xmin, xmax), cmap='gray_r', norm=LogNorm())
s = 'k = '+str(k)
plt.title(s)
plt.xlabel('RA')
plt.ylabel('Dec')
name = 'sgrstream'+str(k)+'.png'
#plt.savefig(name)
plt.show()
#------------------------------------------------------------
# Create the grid on which to evaluate the results
Nx = 50
Ny = 125
xmin, xmax = (-375, -175)
ymin, ymax = (-300, 200)
#------------------------------------------------------------
# Evaluate for several models
Xgrid = np.vstack(map(np.ravel, np.meshgrid(np.linspace(xmin, xmax, Nx),
np.linspace(ymin, ymax, Ny)))).T
kde = KDE(metric='gaussian', h=5)
dens_KDE = kde.fit(X).eval(Xgrid).reshape((Ny, Nx))
knn5 = KNeighborsDensity('bayesian', 5)
dens_k5 = knn5.fit(X).eval(Xgrid).reshape((Ny, Nx))
knn40 = KNeighborsDensity('bayesian', 40)
dens_k40 = knn40.fit(X).eval(Xgrid).reshape((Ny, Nx))
#------------------------------------------------------------
# Plot the results
fig = plt.figure(figsize=(9, 4.0))
fig.subplots_adjust(left=0.1, right=0.95, bottom=0.14, top=0.9,
hspace=0.01, wspace=0.01)
# First plot: scatter the points
ax1 = plt.subplot(221, aspect='equal')
ax1.scatter(X[:, 1], X[:, 0], s=1, lw=0, c='k')
ax1.text(0.98, 0.95, "input", ha='right', va='top',
delimiter=",",
skiprows=1,
usecols=(1, 2, 3, 10, 11, 12, 13, 14, 35, 37, 39, 41, 43, 49, 50, 51)
) #dr7objid, petromag_u, petromag_g, petromag_r, pertomag_i, petromag_z, z(redshift),h alpha ew, h beta ew, OII ew, h delta ew, spiral, elliptical, uncertain
umr = data[:, 3] - data[:, 5] #u-r color
data = np.column_stack((data, umr))
coords = data[:, 1:3]
print coords.shape
print "started knd"
knd = KNeighborsDensity(
"bayesian", 10
) #try using something other than 10 for n_neighbors values, to experiment + optimize
knd.fit(coords)
density = knd.eval(coords)
data[:, 1] = density #!!!!! CHECK, is this still in order??
data = np.delete(data, 2, 1) #(col# 2 , 0/1 for row/col)
print "finished knd"
isSpiral = data[:, 12] #with one col removed!!
#print data[:5, :]
isElliptical = data[:, 13] #corresponds to the elliptical bool value
isUncertain = data[:, 14]
Nx = 50
Ny = 125
xmin, xmax = (-375, -175)
ymin, ymax = (-300, 200)
#------------------------------------------------------------
# Evaluate for several models
Xgrid = np.vstack(
map(np.ravel,
np.meshgrid(np.linspace(xmin, xmax, Nx), np.linspace(ymin, ymax,
Ny)))).T
kde = KDE(metric='gaussian', h=5)
dens_KDE = kde.fit(X).eval(Xgrid).reshape((Ny, Nx))
knn5 = KNeighborsDensity('bayesian', 5)
dens_k5 = knn5.fit(X).eval(Xgrid).reshape((Ny, Nx))
knn40 = KNeighborsDensity('bayesian', 40)
dens_k40 = knn40.fit(X).eval(Xgrid).reshape((Ny, Nx))
#------------------------------------------------------------
# Plot the results
fig = plt.figure(figsize=(5, 2.2))
fig.subplots_adjust(left=0.12,
right=0.95,
bottom=0.2,
top=0.9,
hspace=0.01,
wspace=0.01)
for N, k, subplot in zip(N_values, k_values, subplots):
ax = fig.add_subplot(subplot)
xN = x[:N]
t = np.linspace(-10, 30, 1000)
# Compute density with KDE
if use_sklearn_KDE:
kde = KernelDensity(0.1, kernel='gaussian')
kde.fit(xN[:, None])
dens_kde = np.exp(kde.score_samples(t[:, None]))
else:
kde = KDE('gaussian', h=0.1).fit(xN[:, None])
dens_kde = kde.eval(t[:, None]) / N
# Compute density with Bayesian nearest neighbors
nbrs = KNeighborsDensity('bayesian', n_neighbors=k).fit(xN[:, None])
dens_nbrs = nbrs.eval(t[:, None]) / N
# plot the results
ax.plot(t,
true_pdf(t),
':',
color='black',
zorder=3,
label="Generating Distribution")
ax.plot(xN, -0.005 * np.ones(len(xN)), '|k')
hist(xN,
bins='blocks',
ax=ax,
normed=True,
zorder=1,
fig.subplots_adjust()
N_values = (500, 5000)
subplots = (211, 212)
k_values = (10, 100)
for N, k, subplot in zip(N_values, k_values, subplots):
ax = fig.add_subplot(subplot)
xN = ent2[:N]
t = np.linspace(-10, 30, 1000)
# Compute density with KDE
kde = KDE('gaussian', h=0.1).fit(xN[:, None])
dens_kde = kde.eval(t[:, None]) / N
# Compute density with Bayesian nearest neighbors
nbrs = KNeighborsDensity('bayesian', n_neighbors=k).fit(xN[:, None])
dens_nbrs = nbrs.eval(t[:, None]) / N
# plot the results
#ax.plot(t, true_pdf(t), ':', color='black', zorder=3,
# label="Generating Distribution")
ax.plot(xN, -0.005 * np.ones(len(xN)), '|k', lw=1.5)
hist(xN, bins='blocks', ax=ax, normed=True, zorder=1,
histtype='stepfilled', lw=1.5, color='k', alpha=0.2,
label="Bayesian Blocks")
ax.plot(t, dens_nbrs, '-', lw=2, color='gray', zorder=2,
label="Nearest Neighbors (k=%i)" % k)
ax.plot(t, dens_kde, '-', color='black', zorder=3,
label="Kernel Density (h=0.1)")
# label the plot
"""
Test density estimation techniques
"""
import pytest
import numpy as np
from numpy.testing import assert_allclose
from scipy.stats import norm
from astroML.density_estimation import KNeighborsDensity, GaussianMixture1D
classifiers = [
KNeighborsDensity(method='simple', n_neighbors=250),
KNeighborsDensity(method='bayesian', n_neighbors=250)
]
@pytest.mark.parametrize("clf", classifiers)
def test_1D_density(clf, atol=100):
np.random.seed(0)
dist = norm(0, 1)
X = dist.rvs((5000, 1))
X2 = np.linspace(-5, 5, 10).reshape((10, 1))
true_dens = dist.pdf(X2[:, 0]) * X.shape[0]
clf.fit(X)
dens = clf.eval(X2)
assert_allclose(dens, true_dens, atol=atol)
def test_gaussian1d():
from astroML.datasets import fetch_sdss_specgals
from astroML.density_estimation import KNeighborsDensity
import time
#Trying u-g, g-r, r-i, i-z AS WELL AS raw u, g, r, i, z, still including all 5 spectral lines
#get just spirals and ellipticals in 1 array, shuffle them, then extract the label column
data = np.loadtxt("crossmatched3_combineddata1_srane.txt", delimiter = ",", skiprows = 1, usecols = (1, 2, 3, 49, 50, 51)) #dr7objid, petromag_u, petromag_g, petromag_r, pertomag_i, petromag_z, z(redshift),h alpha ew, h beta ew, OII ew, h delta ew, spiral, elliptical, uncertain
coords = data[:, 1:3]
print coords.shape
knd = KNeighborsDensity("bayesian", 10)
knd.fit(coords)
density = knd.eval(coords)
data[:, 1] = density
data = np.delete(data, 2, 1) #(col# 2 , 0/1 for row/col)
isSpiral = data[:,2]#with one col removed!!
print data[:5, :]
isElliptical = data[:,3] #corresponds to the elliptical bool value
isUncertain = data[:,4]
ellipticals = data[isElliptical == 1]
#fit_table=np.delete(fit_table,np.argwhere(fit_table[:,1] < 0),0)
#fit_table=fit_table[(fit_table[:,0] >= 0)]
#Preparing the grid for evaluation. Nx and Ny can be set separately
N = len(fit)
Ny = N
Nx = N
xmin, xmax = (min(fit[:, 0]), max(fit[:, 0]))
ymin, ymax = (min(fit[:, 1]), max(fit[:, 1]))
x = np.linspace(xmin, xmax, Nx)
y = np.linspace(ymin, ymax, Ny)
b = np.vstack(map(np.ravel, np.meshgrid(x, y))).T
#Density estimation fit and evaluation
k = 40
knn = KNeighborsDensity('bayesian', k)
knn.fit(fit)
c = knn.eval(b).reshape((Ny, Nx))
#Getting optimization data and science data from what's left
others = np.delete(table.data, selection, 0)
N2 = 1000
np.random.seed(1)
optimize = np.random.randint(0, len(others), size=N2)
opt = others[optimize]
others = np.delete(others, optimize, 0)
np.random.seed(2)
science = np.random.randint(0, len(others), size=N2)
