fig.suptitle('JS diverence, clustering = all times')
if fig_dir is not None:
fig.savefig(
os.path.join(fig_dir, 'JSD_time_clustering_tbins=%d.pdf' % (tbins)))
### Manifold embeddings
reload(fplt)
plt.figure(200)
plt.clf()
#label = range(tbins) * nstrains
#label = list(np.sort(range(nstages) * sbins)) * nstrains
label = np.sort(range(nstrains) * tbins)
res = fplt.plot_manifold_embeddings(ent_t['all'],
precomputed=True,
label=label)
reload(fplt)
fig = plt.figure(50)
plt.clf()
label = np.sort(range(nstrains) * tbins)
colors = ['red', 'blue', 'darkgreen', 'orange', 'purple', 'gray']
cmaps = [
plt.cm.Reds, plt.cm.Blues, plt.cm.Greens, plt.cm.Oranges, plt.cm.Purples,
plt.cm.gray_r
]
i = 0
for k, v, in res.iteritems():
plt.subplot(1, 3, i + 1)
i += 1
plt.plot(dat_mean) plt.plot(dat_var) reload(als) reload(fplt) plt.figure(10); plt.clf() fplt.plot_pca(dist.T) plt.figure(11); plt.clf(); fplt.plot_nmf(dist.T, n_components=None) reload(fplt) plt.figure(20); fplt.plot_manifold_embeddings(dist.T, n_components=3, n_neighbors=20) plt.tight_layout() ### Fit distributions #dist_names = ['gamma', 'beta', 'rayleigh', 'norm', 'pareto' import scipy; dist_name = 'beta'; dist_form = getattr(scipy.stats, dist_name) param = np.zeros((dist_form.numargs + 2, nbins)); eps = 10e-9; for b in range(1): dd = dat_bin[:,b].copy();