def plot_feature_rank(dataset, show=False):
print "Plotting plot_feature_rank, dataset: %s" % dataset
dfs = get_feature_rankings(dataset=dataset, polynomial_terms=False)
order = dfs[['feature', 'weight', 'group']].groupby([
'feature', 'group'
])['weight'].mean().reset_index().sort_values(['group', 'weight'],
ascending=[True,
False])['feature']
plot_specs = {
'x_col': 'weight',
'y_col': 'feature',
'hue_col': 'group',
'x_label': 'Score',
'y_label': 'Feature',
'order': order,
'dodge': False,
'labelsize': 8,
'figsize': (8, 11),
'font_scale': 0.8,
'fontsize': 14,
'show': show,
'y_lim': None,
'capsize': .2,
'title': "Feature Importance",
}
figname = 'feature_rank_%s.pdf' % dataset
bar_plot(dfs, figname, **plot_specs)
def domain_adaptation_plot_helper(classifiers, metric='acc'):
print "Plotting domain_adaptation, classifiers: %s" % classifiers
METHODS = ['target_only', 'source_only', 'relabeled', 'augment', 'coral']
dfs = []
for method in METHODS:
for classifier in classifiers:
df = get_da_results(classifier, method, metric)
util.print_ci_from_df(df['folds'], method, classifier)
dfs.append(df)
dfs = pd.concat(dfs)
if metric == 'acc':
y_label = "Accuracy"
elif metric == 'fms':
y_label = "F-Measure"
else:
y_label = "AUC"
plot_specs = {
'x_col': 'method',
'y_col': 'folds',
'hue_col': 'model',
'x_label': 'Model',
'figsize': (10, 8),
'font_scale': 1.2,
'fontsize': 20,
'y_label': y_label,
'y_lim': (0, 1)
}
figname = 'domain_adapt_plot_%s_%s.pdf' % (metric, classifiers[1])
bar_plot(dfs, figname, **plot_specs)
def vanilla_feature_set_plot(show=False):
print "Plotting vanilla_feature_set_plot"
dfs = []
classifiers = models.CLASSIFIER_KEYS
for classifier in classifiers:
for metric in models.METRICS:
df = get_vanilla_results(classifier, metric)
util.print_ci_from_df(df['folds'], classifier, metric)
dfs.append(df)
dfs = pd.concat(dfs)
plot_specs = {
'x_col': 'model',
'y_col': 'folds',
'hue_col': 'metric',
'x_label': 'Model',
'y_label': 'Performance',
'figsize': (12, 10),
'font_scale': 1.2,
'fontsize': 20,
'y_lim': None,
'show': show,
'title': "10-Fold Cross Validation Performance"
}
figname = 'vanilla_results.pdf'
bar_plot(dfs, figname, **plot_specs)
def blog_plot():
print "Plotting blog_plot"
metrics = models.METRICS
dfs = []
for classifier in models.CLASSIFIER_KEYS:
for metric in metrics:
df = get_blog_results(classifier, metric)
util.print_ci_from_df(df['folds'], classifier, metric)
dfs.append(df)
dfs = pd.concat(dfs)
plot_specs = {
'x_col': 'model',
'y_col': 'folds',
'hue_col': 'metric',
'x_label': 'Model',
'y_label': 'Performance',
'font_scale': 1.2,
'fontsize': 20,
'rotation': 15
}
figname = 'blog_plot.pdf'
bar_plot(dfs, figname, **plot_specs)
def plot_blog_feature_rank(show=False):
print "Plotting plot_blog_feature_rank"
dfs = get_blog_feature_rankings()
order = dfs[['feature', 'weight', 'group']].groupby([
'feature', 'group'
])['weight'].mean().reset_index().sort_values(['group', 'weight'],
ascending=[True,
False])['feature']
plot_specs = {
'x_col': 'weight',
'y_col': 'feature',
'hue_col': 'group',
'x_label': 'Feature Score',
'y_label': 'Feature Sets',
'order': order,
'dodge': False,
'labelsize': 8,
'figsize': (8, 11),
'show': show,
'y_lim': None,
'capsize': .2,
}
figname = 'blog_feature_rank.pdf'
bar_plot(dfs, figname, **plot_specs)
def blog_ablation_plot(metric='acc'):
print "Plotting blog_ablation_plot, metric: %s" % metric
classifiers = models.CLASSIFIER_KEYS
ablation_sets = models.BLOG_FEATURE_SETS
classifiers.remove('DummyClassifier')
dfs = []
for classifier in classifiers:
for ab_set in ablation_sets:
df = get_ablation_results(ab_set, classifier, metric,
BLOG_ABLATION_PREFIX)
util.print_ci_from_df(df['folds'], classifier, metric)
dfs.append(df)
dfs = pd.concat(dfs)
human_readable = {"acc": "Accuracy", "fms": "F-Measure", "roc": "AUC"}
plot_specs = {
'x_col': 'ablation_set',
'y_col': 'folds',
'hue_col': 'model',
'x_label': 'Feature Set',
'y_label': "Change in %s " % human_readable[metric],
'title': "Feature Ablation",
'figsize': (10, 8),
'fontsize': 20,
'font_scale': 1.2,
'y_lim': None,
'errwidth': 0.75,
'labelsize': 10,
'rotation': 15
}
figname = 'blog_ablation_plot.pdf'
bar_plot(dfs, figname, **plot_specs)
def barplot():
labels_string = request.args.get('labels')
values_string = request.args.get('values')
xlabel = request.args.get('xlabel')
ylabel = request.args.get('ylabel')
title = request.args.get('title')
try:
labels, values = labels_string.split(','), util.string_to_list(
values_string)
except Exception as e:
print str(e)
return 'Nothing to do here.'
if len(labels) != len(values):
return 'Nothing to do here.'
response = util.bar_plot(labels, values, xlabel, ylabel, title)
return response
def p_Xw_i_MISE(mock, ell=0, rebin=None, krange=None, method='choletsky', b=0.1):
''' Examine the pdf of X_w^i components that deviate significantly from
N(0,1) based on MISE
'''
Pk = NG.dataX(mock, ell=ell, rebin=rebin, krange=krange)
X, _ = NG.meansub(Pk)
X_w, W = NG.whiten(X, method=method) # whitened data
# calculate the chi-squared values of each p(X_w^i)
x = np.arange(-5., 5.1, 0.1)
mise = np.zeros(X_w.shape[1])
for i_bin in range(X_w.shape[1]):
mise[i_bin] = NG.MISE(X_w[:,i_bin], b=b)
# plot the most discrepant components.
prettyplot()
fig = plt.figure()
sub = fig.add_subplot(111)
i_sort = np.argsort(mise)
print 'outlier bins = ', i_sort[-5:]
print 'mise = ', mise[i_sort[-5:]]
nbin = int(10./b)
for i_bin in i_sort[-10:]:
hb_Xi, Xi_edges = np.histogram(X_w[:,i_bin], bins=nbin, range=[-5., 5.], normed=True)
p_X_w_arr = UT.bar_plot(Xi_edges, hb_Xi)
sub.plot(p_X_w_arr[0], p_X_w_arr[1])
sub.plot(x, UT.gauss(x, 1., 0.), c='k', lw=3, label='$\mathcal{N}(0,1)$')
sub.set_xlim([-2.5, 2.5])
sub.set_xlabel('$\mathtt{X^{i}_{W}}$', fontsize=25)
sub.set_ylim([0., 0.6])
sub.set_ylabel('$\mathtt{P(X^{i}_{W})}$', fontsize=25)
sub.legend(loc='upper right')
str_rebin = ''
if rebin is not None:
str_rebin = '.rebin'+str(rebin)
f = ''.join([UT.fig_dir(), 'tests/test.p_Xw_i_outlier.', method, '.', mock, '.ell', str(ell),
str_rebin, '.b', str(b), '.png'])
fig.savefig(f, bbox_inches='tight')
return None
def div_K(div_func='kl'):
''' compare the KL or Renyi divergence for the following with their using different K values
- D( gauss(C_X) || gauss(C_X) )
- D( mock X || gauss(C_X))
- D( mock X || p(X) KDE)
- D( mock X || p(X) GMM)
- D( mock X || PI p(X^i_ICA) KDE)
- D( mock X || PI p(X^i_ICA) GMM)
'''
lbls = [r'$D( P(k) \parallel \mathcal{N}({\bf C}))$',
r'$D( P(k) \parallel p_\mathrm{KDE}(P(k)))$',
r'$D( P(k) \parallel p_\mathrm{GMM}(P(k)))$',
r'$D( P(k) \parallel \prod_i p_\mathrm{KDE}(P(k)_i^\mathrm{ICA}))$',
r'$D( P(k) \parallel \prod_i p_\mathrm{GMM}(P(k)_i^\mathrm{ICA}))$']
fig = plt.figure(figsize=(20,4))
for i_obv, obv in enumerate(['pk.ngc', 'gmf']):
if obv == 'pk.ngc':
Nref = 2000
if div_func == 'kl': hranges = [[-0.5, 0.5], [-0.5, 7.], [-0.5, 0.5], [-0.5, 0.5], [-0.5, 0.5]]##7.]
else: hranges = [[-0.5, 0.5] for i in range(5)]
Ks = [5, 10, 15]
elif obv == 'gmf':
Nref = 10000
hranges = [[-0.1, 0.4], [-0.1, 0.4], [-0.1, 0.4], [-0.1, 0.4], [-0.1, 0.4]]##7.]
Ks = [10]
for K in Ks:
fs = ['pX_gauss.K'+str(K), 'pX_scottKDE.K'+str(K), 'pX_GMM.K'+str(K)+'.ncomp30',
'pXi_ICA_scottKDE.K'+str(K), 'pXi_ICA_GMM.K'+str(K)+'.ncomp30']
divs, divs_ref = [], []
for f in fs:
f_div = ''.join([UT.dat_dir(), 'diverg.', obv, '.', f, '.Nref', str(Nref), '.',
div_func, '.dat'])
try:
div = np.loadtxt(f_div)
except IOError:
print f_div
continue
divs.append(div)
nbins = 50
bkgd = fig.add_subplot(2,1,i_obv+1, frameon=False)
for i_div, div, lbl in zip(range(len(fs)), divs, lbls):
sub = fig.add_subplot(2,5,len(fs)*i_obv+i_div+1)
y_max = 0.
hh = np.histogram(div, normed=True, range=hranges[i_div], bins=nbins)
bp = UT.bar_plot(*hh)
sub.fill_between(bp[0], np.zeros(len(bp[0])), bp[1], edgecolor='none')
y_max = max(y_max, bp[1].max())
sub.set_xlim(hranges[i_div])
sub.set_ylim([0., y_max*1.4])
if i_obv == 0:
sub.set_title(lbl)
if div_func == 'kl':
bkgd.set_xlabel(r'KL divergence', fontsize=20, labelpad=20)
elif div_func == 'renyi0.5':
bkgd.set_xlabel(r'R\'enyi-$\alpha$ divergence', fontsize=20, labelpad=20)
bkgd.set_xticklabels([])
bkgd.set_yticklabels([])
bkgd.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')
fig.subplots_adjust(wspace=.15, hspace=0.3)
f_fig = ''.join([UT.fig_dir(), 'tests/Ktest_kNNdiverg.', div_func, '.png'])
fig.savefig(f_fig, bbox_inches='tight')
return None
def p_Xw_i(mock, ell=0, rebin=None, krange=None, ica=False, pca=False):
''' Test the probability distribution function of each X_w^i
component -- p(X_w^i). First compare the histograms of p(X_w^i)
with N(0,1). Then compare the gaussian KDE of p(X_w^i).
'''
Pk = NG.dataX(mock, ell=ell, rebin=rebin, krange=krange)
X, _ = NG.meansub(Pk)
str_w = 'W'
if ica and pca:
raise ValueError
if ica: # ICA components
# ICA components do not need to be Gaussian.
# in fact the whole point of the ICA transform
# is to capture the non-Gaussianity...
X_white, _ = NG.whiten(X) # whitened data
X_w, _ = NG.Ica(X_white)
str_w = 'ICA'
if pca: # PCA components
X_w, _ = NG.whiten(X, method='pca') # whitened data
str_w = 'PCA'
if not ica and not pca: # just whitened
X_w, W = NG.whiten(X) # whitened data
# p(X_w^i) histograms
fig = plt.figure(figsize=(15,7))
sub = fig.add_subplot(121)
for i_bin in range(X_w.shape[1]):
p_X_w, edges = np.histogram(X_w[:,i_bin], normed=True)
p_X_w_arr = UT.bar_plot(edges, p_X_w)
sub.plot(p_X_w_arr[0], p_X_w_arr[1])
x = np.arange(-5., 5.1, 0.1)
sub.plot(x, UT.gauss(x, 1., 0.), c='k', lw=3, label='$\mathcal{N}(0,1)$')
sub.set_xlim([-2.5, 2.5])
sub.set_xlabel('$\mathtt{X_{'+str_w+'}}$', fontsize=25)
sub.set_ylim([0., 0.6])
sub.set_ylabel('$\mathtt{P(X_{'+str_w+'})}$', fontsize=25)
sub.legend(loc='upper right')
# p(X_w^i) gaussian KDE fits
pdfs = NG.p_Xw_i(X_w, range(X_w.shape[1]), x=x)
sub = fig.add_subplot(122)
for i_bin in range(X_w.shape[1]):
sub.plot(x, pdfs[i_bin])
sub.plot(x, UT.gauss(x, 1., 0.), c='k', lw=3, label='$\mathcal{N}(0,1)$')
sub.set_xlim([-2.5, 2.5])
sub.set_xlabel('$\mathtt{X_{W}}$', fontsize=25)
sub.set_ylim([0., 0.6])
sub.set_ylabel('$\mathtt{P(X_{W})}$', fontsize=25)
sub.legend(loc='upper right')
str_ica, str_pca = '', ''
if ica:
str_ica = '.ICA'
if pca:
str_pca = '.PCA'
if rebin is None:
f = ''.join([UT.fig_dir(), 'tests/test.p_Xw_i', str_pca, str_ica, '.', mock, '.ell', str(ell), '.png'])
else:
f = ''.join([UT.fig_dir(), 'tests/test.p_Xw_i', str_pca, str_ica, '.', mock, '.ell', str(ell), '.rebin', str(rebin), '.png'])
fig.savefig(f, bbox_inches='tight')
return None
def GMF_p_Xw_i(ica=False, pca=False):
''' Test the probability distribution function of each transformed X
component -- p(X^i). First compare the histograms of p(X_w^i)
with N(0,1). Then compare the gaussian KDE of p(X_w^i).
'''
gmf = NG.X_gmf_all() # import all the GMF mocks
X, _ = NG.meansub(gmf)
str_w = 'W'
if ica and pca: raise ValueError
if ica: # ICA components
# ICA components do not need to be Gaussian.
# in fact the whole point of the ICA transform
# is to capture the non-Gaussianity...
X_white, _ = NG.whiten(X) # whitened data
X_w, _ = NG.Ica(X_white)
str_w = 'ICA'
if pca: # PCA components
X_w, _ = NG.whiten(X, method='pca') # whitened data
str_w = 'PCA'
if not ica and not pca: # just whitened
X_w, W = NG.whiten(X) # whitened data
# p(X_w^i) histograms
fig = plt.figure(figsize=(5*gmf.shape[1],4))
for icomp in range(gmf.shape[1]):
sub = fig.add_subplot(1, gmf.shape[1], icomp+1)
# histogram of X_w^i s
hh = np.histogram(X_w[:,icomp], normed=True, bins=50, range=[-5., 5.])
p_X_w_arr = UT.bar_plot(*hh)
sub.fill_between(p_X_w_arr[0], np.zeros(len(p_X_w_arr[1])), p_X_w_arr[1],
color='k', alpha=0.25)
x = np.linspace(-5., 5., 100)
sub.plot(x, UT.gauss(x, 1., 0.), c='k', lw=2, ls=':', label='$\mathcal{N}(0,1)$')
# p(X_w^i) gaussian KDE fits
t_start = time.time()
pdf = NG.p_Xw_i(X_w, icomp, x=x, method='gkde')
sub.plot(x, pdf, lw=2.5, label='Gaussian KDE')
print 'scipy Gaussian KDE ', time.time()-t_start
# p(X_w^i) SKlearn KDE fits
t_start = time.time()
pdf = NG.p_Xw_i(X_w, icomp, x=x, method='sk_kde')
sub.plot(x, pdf, lw=2.5, label='SKlearn KDE')
print 'SKlearn CV best-fit KDE ', time.time()-t_start
# p(X_w^i) statsmodels KDE fits
t_start = time.time()
pdf = NG.p_Xw_i(X_w, icomp, x=x, method='sm_kde')
sub.plot(x, pdf, lw=2.5, label='StatsModels KDE')
print 'Stats Models KDE ', time.time()-t_start
# p(X_w^i) GMM fits
pdf = NG.p_Xw_i(X_w, icomp, x=x, method='gmm', n_comp_max=20)
sub.plot(x, pdf, lw=2.5, ls='--', label='GMM')
sub.set_xlim([-3., 3.])
sub.set_xlabel('$X_{'+str_w+'}^{('+str(icomp)+')}$', fontsize=25)
sub.set_ylim([0., 0.6])
if icomp == 0:
sub.set_ylabel('$P(X_{'+str_w+'})$', fontsize=25)
sub.legend(loc='upper left', prop={'size': 15})
str_ica, str_pca = '', ''
if ica: str_ica = '.ICA'
if pca: str_pca = '.PCA'
f = ''.join([UT.fig_dir(), 'tests/test.GMF_p_Xw_i', str_pca, str_ica, '.png'])
fig.savefig(f, bbox_inches='tight')
return None
def divGMF(div_func='kl', Nref=1000, K=5, n_mc=10, n_comp_max=10, n_mocks=2000):
''' compare the divergence estimates between
D( gauss(C_gmf) || gauss(C_gmf) ), D( gmfs || gauss(C_gmf) ),
D( gmfs || p(gmfs) KDE), D( gmfs || p(gmfs) GMM),
D( gmfs || PI p(gmfs^i_ICA) KDE), and D( gmfs || PI p(gmfs^i_ICA) GMM)
'''
if isinstance(Nref, float):
Nref = int(Nref)
# read in mock GMFs from all HOD realizations (20,000 mocks)
gmfs_mock = NG.X_gmf_all()[:n_mocks]
n_mock = gmfs_mock.shape[0] # number of mocks
print("%i mocks" % n_mock)
gmfs_mock_meansub, _ = NG.meansub(gmfs_mock) # mean subtract
X_w, W = NG.whiten(gmfs_mock_meansub)
X_ica, _ = NG.Ica(X_w) # ICA transformation
C_gmf = np.cov(X_w.T) # covariance matrix
# p(gmfs) GMM
gmms, bics = [], []
for i_comp in range(1,n_comp_max+1):
gmm = GMix(n_components=i_comp)
gmm.fit(X_w)
gmms.append(gmm)
bics.append(gmm.bic(X_w))
ibest = np.array(bics).argmin()
kern_gmm = gmms[ibest]
# p(gmfs) KDE
t0 = time.time()
grid = GridSearchCV(skKDE(),
{'bandwidth': np.linspace(0.1, 1.0, 30)},
cv=10) # 10-fold cross-validation
grid.fit(X_w)
kern_kde = grid.best_estimator_
dt = time.time() - t0
print('%f sec' % dt)
# PI p(gmfs^i_ICA) GMM
kern_gmm_ica = []
for ibin in range(X_ica.shape[1]):
gmms, bics = [], []
for i_comp in range(1,n_comp_max+1):
gmm = GMix(n_components=i_comp)
gmm.fit(X_ica[:,ibin][:,None])
gmms.append(gmm)
bics.append(gmm.bic(X_ica[:,ibin][:,None]))
ibest = np.array(bics).argmin()
kern_gmm_ica.append(gmms[ibest])
# PI p(gmfs^i_ICA) KDE
kern_kde_ica = []
for ibin in range(X_ica.shape[1]):
t0 = time.time()
grid = GridSearchCV(skKDE(),
{'bandwidth': np.linspace(0.1, 1.0, 30)},
cv=10) # 10-fold cross-validation
grid.fit(X_ica[:,ibin][:,None])
kern_kde_ica.append(grid.best_estimator_)
dt = time.time() - t0
print('%f sec' % dt)
# caluclate the divergences now
div_gauss_ref, div_gauss = [], []
div_gmm, div_gmm_ica = [], []
div_kde, div_kde_ica = [], []
for i in range(n_mc):
print('%i montecarlo' % i)
t_start = time.time()
# reference divergence in order to showcase the estimator's scatter
# Gaussian distribution described by C_gmf with same n_mock mocks
gauss = mvn(np.zeros(gmfs_mock.shape[1]), C_gmf, size=n_mock)
div_gauss_ref_i = NG.kNNdiv_gauss(gauss, C_gmf, Knn=K, div_func=div_func, Nref=Nref)
div_gauss_ref.append(div_gauss_ref_i)
# estimate divergence between gmfs_white and a
# Gaussian distribution described by C_gmf
div_gauss_i = NG.kNNdiv_gauss(X_w, C_gmf, Knn=K, div_func=div_func, Nref=Nref)
div_gauss.append(div_gauss_i)
# D( gmfs || p(gmfs) GMM)
div_gmm_i = NG.kNNdiv_Kernel(X_w, kern_gmm, Knn=K, div_func=div_func,
Nref=Nref, compwise=False)
div_gmm.append(div_gmm_i)
# D( gmfs || p(gmfs) KDE)
div_kde_i = NG.kNNdiv_Kernel(X_w, kern_kde, Knn=K, div_func=div_func,
Nref=Nref, compwise=False)
div_kde.append(div_kde_i)
# D( gmfs || PI p(gmfs^i_ICA) GMM),
div_gmm_ica_i = NG.kNNdiv_Kernel(X_ica, kern_gmm_ica, Knn=K, div_func=div_func,
Nref=Nref, compwise=True)
div_gmm_ica.append(div_gmm_ica_i)
# D( gmfs || PI p(gmfs^i_ICA) KDE),
div_kde_ica_i = NG.kNNdiv_Kernel(X_ica, kern_kde_ica, Knn=K, div_func=div_func,
Nref=Nref, compwise=True)
div_kde_ica.append(div_kde_ica_i)
print('t= %f sec' % round(time.time()-t_start,2))
fig = plt.figure(figsize=(10,5))
sub = fig.add_subplot(111)
hrange = [-0.15, 0.6]
nbins = 50
divs = [div_gauss_ref, div_gauss, div_gmm, div_kde, div_gmm_ica, div_kde_ica]
labels = ['Ref.', r'$D(\{\zeta_i^{(m)}\}\parallel \mathcal{N}({\bf C}^{(m)}))$',
r'$D(\{\zeta^{(m)}\}\parallel p_\mathrm{GMM}(\{\zeta^{m}\}))$',
r'$D(\{\zeta^{(m)}\}\parallel p_\mathrm{KDE}(\{\zeta^{m}\}))$',
r'$D(\{\zeta_\mathrm{ICA}^{(m)}\}\parallel \prod_{i} p^\mathrm{GMM}(\{\zeta_{i, \mathrm{ICA}}^{m}\}))$',
r'$D(\{\zeta_\mathrm{ICA}^{(m)}\}\parallel \prod_{i} p^\mathrm{KDE}(\{\zeta_{i, \mathrm{ICA}}^{m}\}))$']
y_max = 0.
for div, lbl in zip(divs, labels):
hh = np.histogram(np.array(div), normed=True, range=hrange, bins=nbins)
bp = UT.bar_plot(*hh)
sub.fill_between(bp[0], np.zeros(len(bp[0])), bp[1], edgecolor='none',
alpha=0.5, label=lbl)
y_max = max(y_max, bp[1].max())
if (np.average(div) < hrange[0]) or (np.average(div) > hrange[1]):
print('divergence of %s (%f) is outside range' % (lbl, np.average(div)))
sub.set_xlim(hrange)
sub.set_ylim([0., y_max*1.2])
sub.legend(loc='upper left', prop={'size': 15})
# xlabels
if 'renyi' in div_func:
alpha = float(div_func.split(':')[-1])
sub.set_xlabel(r'Renyi-$\alpha='+str(alpha)+'$ divergence', fontsize=20)
elif 'kl' in div_func:
sub.set_xlabel(r'KL divergence', fontsize=20)
if 'renyi' in div_func: str_div = 'renyi'+str(alpha)
elif div_func == 'kl': str_div = 'kl'
f_fig = ''.join([UT.fig_dir(), 'tests/kNN_divergence.gmf.K', str(K), '.', str(n_mocks),
'.', str_div, '.png'])
fig.savefig(f_fig, bbox_inches='tight')
return None
def div_ICA(obv='pk.ngc', K=10, div_func='kl'):
''' compare the KL or Renyi divergence for different ICA decomposition algorithms
FastICA deflation, FastICA parallel, Infomax ICA
- D( mock X || PI p(X^i_ICA) KDE)
- D( mock X || PI p(X^i_ICA) GMM)
'''
if obv == 'pk.ngc': str_obv = 'P(k)'
elif obv == 'gmf': str_obv = '\zeta(N)'
lbls = [r'$D( '+str_obv+' \parallel \prod_i p_\mathrm{KDE}(P(k)_i^\mathrm{ICA}))$',
r'$D( '+str_obv+' \parallel \prod_i p_\mathrm{GMM}(P(k)_i^\mathrm{ICA}))$']
icas = ['ICA', 'parICA']
if obv == 'pk.ngc':
Nref = 2000
hrange = [-0.5, 0.5]
elif obv == 'gmf':
Nref = 10000
hranges = [-0.1, 0.4]
fig = plt.figure(figsize=(10,4))
bkgd = fig.add_subplot(111, frameon=False)
for i_div, str_div in enumerate(['scottKDE.K'+str(K), 'GMM.K'+str(K)+'.ncomp30']):
divs = []
for ica in icas:
f_div = ''.join([UT.dat_dir(), 'diverg.', obv,
'.pXi_', ica, '_', str_div, '.Nref', str(Nref), '.', div_func, '.dat'])
try:
div = np.loadtxt(f_div)
except IOError:
print f_div
continue
divs.append(div)
nbins = 50
sub = fig.add_subplot(1,2,i_div+1)
y_max = 0.
for div, ica in zip(divs, icas):
print np.mean(div)
hh = np.histogram(div, normed=True, range=hrange, bins=nbins)
bp = UT.bar_plot(*hh)
sub.fill_between(bp[0], np.zeros(len(bp[0])), bp[1], edgecolor='none', label=ica)
y_max = max(y_max, bp[1].max())
if i_div == 0: sub.legend(loc='upper left', prop={'size': 20})
sub.set_xlim(hrange)
sub.set_ylim([0., y_max*1.4])
sub.set_title(lbls[i_div])
if div_func == 'kl':
bkgd.set_xlabel(r'KL divergence', fontsize=20, labelpad=20)
elif div_func == 'renyi0.5':
bkgd.set_xlabel(r'R\'enyi-$\alpha$ divergence', fontsize=20, labelpad=20)
bkgd.set_xticklabels([])
bkgd.set_yticklabels([])
bkgd.tick_params(labelcolor='none', top='off', bottom='off', left='off', right='off')
fig.subplots_adjust(wspace=.15, hspace=0.3)
f_fig = ''.join([UT.fig_dir(), 'tests/',
'ICA_kNNdiverg.', obv, '.K', str(K), '.', div_func, '.png'])
fig.savefig(f_fig, bbox_inches='tight')
return None
# Plotted read functions include:
# - readConsent
# - readArtwork
# - getHistoryForArtwork
# - readLog
# - readModel
# - readBalance
# - invokeModel
import util
from matplotlib import pyplot as plt
if __name__ == "__main__":
filterFunc = [
'initialConsent', 'grantRevokeConsent', 'uploadArtwork',
'transferArtwork', 'addLog', 'addModel', 'addWallet',
'transferBlalance'
]
data = util.concat_result(filterFunc, 7) # 7 for tps
fig, ax = plt.subplots(figsize=(10, 6))
util.bar_plot(ax, data, total_width=.75, single_width=.9)
plt.xticks(range(7), ["10", "20", "30", "40", "50", "60", "70"])
plt.xlabel('txDuration (sec)')
plt.ylabel('Throughtput (tps)')
plt.title(
'Read performance of different functions under different transaction duration'
)
plt.show()
def new_feature_set_plot(metric='acc', absolute=True, poly=True, show=False):
print "Plotting new_feature_set_plot, metric: %s" % metric
classifiers = list(models.CLASSIFIER_KEYS)
new_features = []
if absolute:
new_features += ['none']
new_features += models.NEW_FEATURE_SETS
classifiers.remove('DummyClassifier')
dfs = []
for fs in new_features:
for classifier in classifiers:
df = get_new_feature_results(fs,
classifier,
metric,
absolute=absolute,
poly=poly)
util.print_ci_from_df(df['folds'], fs, classifier)
dfs.append(df)
dfs = pd.concat(dfs)
dfs = dfs.replace('none', 'baseline')
y_lim = (.68, .90)
if metric == 'acc':
y_label = "Accuracy"
elif metric == 'fms':
y_label = "F-Measure"
else:
y_label = "AUC"
y_lim = (.70, .95)
figname = 'new_feature_plot_%s' % metric
title = 'Performance w/ New Feature Sets'
if not absolute:
y_label = "Change in %s" % y_label
y_lim = (-.10, .10)
figname = figname + '_relative'
title = 'Change in Performance w/ New Feature Sets'
plot_specs = {
'x_col': 'new_feature_set',
'y_col': 'folds',
'hue_col': 'model',
'x_label': 'Feature Set',
'y_label': y_label,
'y_lim': y_lim,
'figsize': (10, 8),
'fontsize': 20,
'font_scale': 1.2,
'labelsize': 15,
'show': show,
'title': title,
}
# We use polynomial terms as well for halves
if poly:
dfs = dfs.replace('halves', 'halves+quadratic')
else:
figname = figname + '_without_quadratic'
figname = figname + '.pdf'
bar_plot(dfs, figname, **plot_specs)
def groupcatSFMS(mrange=[10.6,10.8]):
'''Figure of the z~0 group catalog.
Panel a) SFR-M* relation
Panel b) P(SSFR) with SFMS fitting
'''
# Read in Jeremy's group catalog with Mr_cut = -18
gc = Cat.Observations('group_catalog', Mrcut=18, position='central')
gc_cat = gc.Read()
fig = plt.figure(figsize=(10,5))
# fit the SFMS using lettalkaboutquench sfms fitting
_fSFMS = fstarforms()
_fit_logm, _fit_logsfr = _fSFMS.fit(gc_cat['mass'], gc_cat['sfr'], method='gaussmix', fit_range=None)
logsfr_ms = _fSFMS.powerlaw(logMfid=10.5)
print _fSFMS._powerlaw_m
print _fSFMS._powerlaw_c
fSFMS = fstarforms()
fit_logm, _ = fSFMS.fit(gc_cat['mass'], gc_cat['sfr'], method='gaussmix', fit_range=mrange)
_, fit_fsfms = fSFMS.frac_SFMS()
i_fit = np.abs(fit_logm - np.mean(mrange)).argmin()
# log SFR - log M* highlighting where the SFMS lies
sub1 = fig.add_subplot(1,2,1)
DFM.hist2d(gc_cat['mass'], gc_cat['sfr'], color='#ee6a50',
levels=[0.68, 0.95], range=[[9., 12.], [-3.5, 1.5]],
plot_datapoints=True, fill_contours=False, plot_density=True, ax=sub1)
gc = Cat.Observations('group_catalog', Mrcut=18, position='central')
gc_cat = gc.Read()
#sub1.vlines(mrange[0], -5., 2., color='k', linewidth=2, linestyle='--')
#sub1.vlines(mrange[1], -5., 2., color='k', linewidth=2, linestyle='--')
#sub1.fill_between(mrange, [2.,2.], [-5.,-5], color='#1F77B4', alpha=0.25)
sub1.fill_between(mrange, [2.,2.], [-5.,-5], color='k', linewidth=0, alpha=0.25)
print _fit_logm, _fit_logsfr
sub1.plot(np.linspace(9.8, 11., 10), logsfr_ms(np.linspace(9.8, 11., 10)), c='k', linestyle='--')
sub1.set_xticks([9., 10., 11., 12.])
sub1.set_xlabel('log$(\; M_*\; [M_\odot]\;)$', fontsize=20)
sub1.set_yticks([-3., -2., -1., 0., 1.])
sub1.set_ylabel('log$(\; \mathrm{SFR}\; [M_\odot/\mathrm{yr}]\;)$', fontsize=20)
sub1.text(0.95, 0.1, 'SDSS central galaxies',
ha='right', va='center', transform=sub1.transAxes, fontsize=20)
# P(log SSFR)
sub2 = fig.add_subplot(1,2,2)
inmbin = np.where((gc_cat['mass'] > mrange[0]) & (gc_cat['mass'] < mrange[1]))
bedge, pp = np.histogram(gc_cat['ssfr'][inmbin], range=[-14., -9.], bins=32, normed=True)
pssfr = UT.bar_plot(pp, bedge)
sub2.plot(pssfr[0], pssfr[1], c='k', lw=2)
# overplot GMM component for SFMS
gmm_weights = fSFMS._gmix_weights[i_fit]
gmm_means = fSFMS._gmix_means[i_fit]
gmm_vars = fSFMS._gmix_covariances[i_fit]
icomp = gmm_means.argmax()
xx = np.linspace(-14., -9, 100)
sub2.fill_between(xx, np.zeros(len(xx)), gmm_weights[icomp]*MNorm.pdf(xx, gmm_means[icomp], gmm_vars[icomp]),
color='#1F77B4', linewidth=0)
for i_comp in range(len(gmm_vars)):
if i_comp == 0:
gmm_tot = gmm_weights[i_comp]*MNorm.pdf(xx, gmm_means[i_comp], gmm_vars[i_comp])
else:
gmm_tot += gmm_weights[i_comp]*MNorm.pdf(xx, gmm_means[i_comp], gmm_vars[i_comp])
#sub2.plot(xx, gmm_tot, color='r', linewidth=2)
sub2.set_xlim([-13.25, -9.5])
sub2.set_xticks([-10., -11., -12., -13.][::-1])
#sub2.set_xlim([-9.5, -13.25])
#sub2.set_xticks([-10., -11., -12., -13.])
sub2.set_xlabel('log$(\; \mathrm{SSFR}\; [\mathrm{yr}^{-1}]\;)$', fontsize=20)
sub2.set_ylim([0., 1.5])
sub2.set_yticks([0., 0.5, 1., 1.5])
sub2.set_ylabel('$p\,(\;\mathrm{log}\; \mathrm{SSFR}\;)$', fontsize=20)
# mass bin
sub2.text(0.5, 0.9, '$'+str(mrange[0])+'< \mathrm{log}\, M_* <'+str(mrange[1])+'$',
ha='center', va='center', transform=sub2.transAxes, fontsize=20)
sub2.text(0.9, 0.33, '$f_\mathrm{SFMS}='+str(round(fit_fsfms[i_fit],2))+'$',
ha='right', va='center', transform=sub2.transAxes, fontsize=20)
fig.subplots_adjust(wspace=.3)
fig.savefig(''.join([UT.tex_dir(), 'figs/groupcat.pdf']), bbox_inches='tight', dpi=150)
plt.close()
return None
