def set_up(self):
# load data using kameleon-mcmc code
logger.info("Loading data")
X, y = GlassPosterior._load_glass_data()
# normalise and whiten dataset, as done in kameleon-mcmc code
logger.info("Whitening data")
X -= np.mean(X, 0)
L = np.linalg.cholesky(np.cov(X.T))
X = sp.linalg.solve_triangular(L, X.T, lower=True).T
# build target, as in kameleon-mcmc code
self.gp_posterior = PseudoMarginalHyperparameters(X, y,
self.n_importance,
self.prior_log_pdf,
self.ridge,
num_shogun_threads=1)
def set_up(self):
# load data using kameleon-mcmc code
logger.info("Loading data")
X, y = GlassPosterior._load_glass_data()
# normalise and whiten dataset, as done in kameleon-mcmc code
logger.info("Whitening data")
X -= np.mean(X, 0)
L = np.linalg.cholesky(np.cov(X.T))
X = sp.linalg.solve_triangular(L, X.T, lower=True).T
# build target, as in kameleon-mcmc code
self.gp_posterior = PseudoMarginalHyperparameters(X,
y,
self.n_importance,
self.prior_log_pdf,
self.ridge,
num_shogun_threads=1)
def _load_glass_data(data_dir=os.sep.join(
[os.path.expanduser('~'), "data"])):
filename = os.sep.join([data_dir, "glass.data"])
try:
data = np.loadtxt(filename, delimiter=",")
except IOError:
# make sure dir exists
try:
os.makedirs(data_dir)
except OSError:
pass
url = "https://archive.ics.uci.edu/ml/machine-learning-databases/glass/glass.data"
logger.warning("%s not found. Trying to download from %s" %
(filename, url))
urllib.urlretrieve(url, filename)
# try again
try:
data = np.loadtxt(filename, delimiter=",")
except IOError:
raise RuntimeError(
"Download failed. Please download manually.")
# make sure file is as expected
s_reference = "eb292f3709b6fbbeb18a34f95e2293470cbe58ed"
logger.info("Asserting sha1sum(%s)==%s" % (filename, s_reference))
s = sha1sum(filename)
if s != s_reference:
raise RuntimeError("sha1sum(%s) is %s while reference is %s" %
(filename, s, s_reference))
# create a binary "window glass" vs "non-window glass" labelling
lab = data[:, -1]
lab = np.array([1. if x <= 4 else -1.0 for x in lab])
# cut off ids and labeling
data = data[:, 1:-1]
return data, lab
def _load_glass_data(data_dir=os.sep.join([os.path.expanduser('~'), "data"])):
filename = os.sep.join([data_dir, "glass.data"])
try:
data = np.loadtxt(filename, delimiter=",")
except IOError:
# make sure dir exists
try:
os.makedirs(data_dir)
except OSError:
pass
url = "https://archive.ics.uci.edu/ml/machine-learning-databases/glass/glass.data"
logger.warning("%s not found. Trying to download from %s" % (filename, url))
urllib.urlretrieve (url, filename)
# try again
try:
data = np.loadtxt(filename, delimiter=",")
except IOError:
raise RuntimeError("Download failed. Please download manually.")
# make sure file is as expected
s_reference = "eb292f3709b6fbbeb18a34f95e2293470cbe58ed"
logger.info("Asserting sha1sum(%s)==%s" % (filename, s_reference))
s = sha1sum(filename)
if s != s_reference:
raise RuntimeError("sha1sum(%s) is %s while reference is %s" % (filename,s, s_reference))
# create a binary "window glass" vs "non-window glass" labelling
lab = data[:, -1]
lab = np.array([1. if x <= 4 else -1.0 for x in lab])
# cut off ids and labeling
data = data[:, 1:-1]
return data, lab
