import numpy as np
import matplotlib.pyplot as plt
get_ipython().run_line_magic('matplotlib', 'inline')
import seaborn
seaborn.set_context('talk')
def plot_dataset(dataset, N_per_sample=1, nbins=30):
orig_dataset = dataset.copy()
dataset = dataset.reshape((len(dataset) // N_per_sample, N_per_sample))
dataset = np.mean(dataset, axis=1)
xlim = (6, 14)
fig, ax = plt.subplots(2, 1, figsize=(8, 6), sharex=True)
fig.subplots_adjust(hspace=0)
plt.sca(ax[0])
plt.plot(dataset, np.arange(len(dataset)), lw=1)
plt.title('Dataset')
plt.ylabel('Trial')
plt.sca(ax[1])
bins = np.linspace(*xlim, num=nbins)
plt.hist(dataset, bins=bins, edgecolor='k')
plt.hist(orig_dataset, bins=bins, edgecolor='k', alpha=0.3, zorder=0)
plt.axvline(np.mean(dataset), c='r', label='mean')
plt.axvline(np.mean(dataset) - np.std(dataset), c='r', ls='--', label='stdev')
plt.axvline(np.mean(dataset) + np.std(dataset), c='r', ls='--')
plt.xlim(xlim)
plt.legend(loc='upper right', fontsize='small')
if N_per_sample > 1:
plt.text(xlim[0] * 1.1, plt.ylim()[1] * 0.8,
r'$1/\sqrt{N_s}=' + f'{1/np.sqrt(N_per_sample):.2f}$',
fontsize='small')
plt.xlabel('Mean')
plt.ylabel('Occurence')
return datasetDA / 大一