def plot_overlap_cumulative(df, size=100):
fig, ax = plots.cumulative_distribution(df.w,
df.ovrl,
label='Within Company Overlap')
fig, ax = plots.cumulative_distribution(df.w,
df.e_ovrl,
label='Full network',
fig=fig,
ax=ax)
ax.set_xlabel(r'$P_{>}(w)$')
ax.set_ylabel(r'$\langle O | P_{>}(w) \rangle$')
ax.set_title(
'Overlap as a function of cumulative distribution of number of calls')
ax.legend(loc=0)
fig.savefig(run_path + 'overlap_cumulative_calls.png')
return fig, ax
def plot_btrns_cumulative(df, size=100):
idx = df.c_brtrn.notnull()
fig, ax = plots.cumulative_distribution(df.c_brtrn[idx],
df.ovrl[idx],
size=100)
fig, ax = plots.cumulative_distribution(df.c_wkn_t[idx],
df.ovrl[idx],
size=100,
fig=fig,
ax=ax)
ax.set_xlabel(r'$P_>(\tau_r^{-1})$')
ax.set_ylabel(r'$\langle O | P_>(\tau_r^{-1}) \rangle$')
ax.set_title(
'Overlap as a function of cumulative \n inverse residual waiting time')
ax.legend()
fig.savefig(run_path + 'brtrn_cumulative.png')
return fig, ax
def plot_iet_mean_cumulative(df):
idx = df.c_iet_mu_na.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_mu_na[idx],
df.ovrl[idx],
label='Naive')
idx = df.c_iet_mu_km.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_mu_km[idx],
df.ovrl[idx],
label='KM',
fig=fig,
ax=ax)
ax.legend()
ax.set_xlabel(r'$P_{>}(\tau)$')
ax.set_ylabel(r'$\langle O | P_{>}(\tau) \rangle$')
ax.set_title('Overlap as a function of cumulative mean inter-event time ')
fig.savefig(run_path + 'mean_iet_cumulative.png')
return fig, ax
def plot_iet_rfresh_cumulative(df, factor=1.1):
idx = df.c_iet_rfsh_na.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_rfsh_na[idx],
df.ovrl[idx])
ax.set_xlabel(r'$P_>(f)$')
ax.set_ylabel(r'$\langle O | P_>(f) \rangle$')
ax.set_title('Overlap as a function of cumulative \n relative freshness')
fig.savefig(run_path + 'rfsh_iets_cumulative.png')
return fig, ax
def plot_iet_sig_cumulative(df, factor=1.1):
idx = df.c_iet_sig_na.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_sig_na[idx],
df.ovrl[idx],
label='Naive')
idx = df.c_iet_sig_km.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_sig_km[idx],
df.ovrl[idx],
label='KM',
fig=fig,
ax=ax)
ax.legend()
ax.set_xlabel(r'$P_>(\sigma_{\tau})$')
ax.set_ylabel(r'$\langle O | P_>(\sigma_{\tau}) \rangle$')
ax.set_title(
'Overlap as a function of cumulative \n inter-event time standard deviation'
)
fig.savefig(run_path + 'std_iets_cumulative.png')
return fig, ax
def plot_iet_aget_cumulative(df, size=100):
idx = df.c_iet_age_na.notnull()
fig, ax = plots.cumulative_distribution(1 / df.c_iet_age_na[idx],
df.ovrl[idx],
size=100)
ax.set_xlabel(r'$P_>(\tau_r^{-1})$')
ax.set_ylabel(r'$\langle O | P_>(\tau_r^{-1}) \rangle$')
ax.set_title(
'Overlap as a function of cumulative \n inverse residual waiting time')
fig.savefig(run_path + 'age_iets_cumulative.png')
return fig, ax
def plot_iet_bur_cumulative(df, size=20, arg='-'):
idx = df.c_iet_bur_na.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_bur_na[idx],
df.ovrl[idx],
label='Naive',
size=size,
arg=arg)
idx = df.c_iet_bur_km.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_bur_km[idx],
df.ovrl[idx],
label='KM',
fig=fig,
ax=ax,
size=size,
arg=arg)
idx = df.c_iet_bur_c_na.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_bur_c_na[idx],
df.ovrl[idx],
label='C-Naive',
fig=fig,
ax=ax,
size=size,
arg=arg)
idx = df.c_iet_bur_c_km.notnull()
fig, ax = plots.cumulative_distribution(df.c_iet_bur_c_km[idx],
df.ovrl[idx],
label='C-KM',
fig=fig,
ax=ax,
size=size,
arg=arg)
ax.legend()
ax.set_xlabel(r'$P_{>}(B)$')
ax.set_ylabel(r'$\langle O | P_{>}(B) \rangle$')
ax.set_title('Overlap as a function of cumulative burstiness')
fig.savefig(run_path + 'bur_iet_cumulative.png')
return fig, ax