NUM_PARTS = 16
# make equal-sized partitions corresponding to approximately equal sized age bins
age_bin2tokens_ = make_age_bin2data(CORPUS_NAME)
age_bin2tokens = make_age_bin2data_with_min_size(age_bin2tokens_)
num_bins = len(age_bin2tokens)
WSPACE = 0.0
HSPACE = 0.0
WPAD = 0.0
HPAD = 0.0
PAD = 0.2
y = []
for part in age_bin2tokens.values():
y.append(mtld(part))
# fig
_, ax = plt.subplots(figsize=(6, 4), dpi=configs.Fig.dpi)
ax.set_ylabel('Measure of\nLexical Textual Diversity',
fontsize=configs.Fig.ax_fontsize)
ax.set_xlabel('Partition', fontsize=configs.Fig.ax_fontsize)
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.tick_params(axis='both', which='both', top=False, right=False)
# plot
x = np.arange(num_bins) + 1
ax.plot(x, y, '-')
plot_best_fit_line(ax, x, y, x_pos=0.70, y_pos=0.1)
plt.show()
plt.title(f'{CORPUS_NAME}\nnum_types={num_types:,}, part {part_id + 1} of {num_bins}')
ax.set_xlabel('mean')
ax.set_ylabel('std')
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.tick_params(axis='both', which='both', top=False, right=False)
# plot
ax.text(x=1.0, y=0.3, s='Taylor\'s exponent: {:.3f}'.format(alpha))
ax.loglog(x, y, '.', markersize=2)
ax.loglog(x, amp * (x ** alpha) + 0, '.', markersize=2)
plt.show()
taylor_exponents.append(alpha)
# fig
_, ax = plt.subplots(figsize=(6, 4), dpi=configs.Fig.dpi)
ax.set_ylabel('Taylor Exponent', fontsize=configs.Fig.ax_fontsize)
ax.set_xlabel('Corpus Partition', fontsize=configs.Fig.ax_fontsize)
ax.spines['right'].set_visible(False)
ax.spines['top'].set_visible(False)
ax.tick_params(axis='both', which='both', top=False, right=False)
# plot
x = np.arange(num_bins) + 1
ax.plot(x, taylor_exponents, '-')
plot_best_fit_line(ax, x, taylor_exponents, x_pos=0.70, y_pos=0.1)
plt.show()