#########################################
#Fit scales to logistic function
L_0 = max(scales)
k_0 = 1
x_0 = 1995
L_MLE_scale, k_MLE_scale, x_MLE_scale = util.logistic_est(scales, L_0, k_0, x_0, years, smooth, datatype='fertility', param='Scale', flip=True)
scale_params = L_MLE_scale, k_MLE_scale, x_MLE_scale, np.min(scales)
#Transition graphs
util.plot_data_transition_gen_gamma_estimates(beta_params, alpha_params, m_params, scale_params, start, end, ages, smooth, datatype='fertility')
util.plot_data_transition(fert_data, start, end, ages, smooth, datatype='fertility')
util.plot_data_transition_gen_gamma_overlay_estimates(fert_data, beta_params, alpha_params, m_params, scale_params, start, end, ages, smooth, datatype='fertility')
#Graph comparison between 2014 and 2100
util.plot_2100(beta_params, alpha_params, m_params, scale_params, ages, smooth, datatype='fertility')
##############################################
alphas = []
betas = []
ms = []
scales = []
for year in range(start, end + 1):
#Take 'smooth' years rolling average
mort_yr = util.rolling_avg_year(mort_data, year, smooth)
alpha, beta, m, scale = util.gen_gamma_est(mort_yr, year, smooth, datatype='mortality')
alphas.append(alpha)
betas.append(beta)
ages,
smooth,
datatype='fertility')
util.plot_data_transition(fert_data,
start,
end,
ages,
smooth,
datatype='fertility')
util.overlay_estimates(fert_data,
a_params,
b_params,
p_params,
q_params,
scale_params,
start,
end,
ages,
smooth,
datatype='fertility')
#Graph comparison between 2014 and 2100
util.plot_2100(a_params,
b_params,
p_params,
q_params,
scale_params,
ages,
smooth,
datatype='fertility')
k_0 = 1.5 x_0 = 1995 L_MLE_alpha, k_MLE_alpha, x_MLE_alpha = util.logistic_est(alphas, L_0, k_0, x_0, years, smooth, datatype='population', param='Alpha') alpha_params = L_MLE_alpha, k_MLE_alpha, x_MLE_alpha, np.min(alphas) ######################################### #Fit ms to logistic function L_0 = 5#max(ms) k_0 = 0.2#1e-50 x_0 = 1995 L_MLE_m, k_MLE_m, x_MLE_m = util.logistic_est(ms, L_0, k_0, x_0, years, smooth, datatype='population', param='M') m_params = L_MLE_m, k_MLE_m, x_MLE_m, np.min(ms) ######################################### #Fit scales to log function a_0 = 2e6 b_0 = np.min(years) - 1 c_0 = 1.9 d_0 = np.min(scales) e_0 = 1 a_MLE_scale, b_MLE_scale, c_MLE_scale, d_MLE_scale, e_MLE_scale = util.poly_est(scales, a_0, b_0, c_0, d_0, e_0, years, smooth, datatype='population', param='Scale', print_params=True) scale_params = a_MLE_scale, b_MLE_scale, c_MLE_scale, d_MLE_scale, e_MLE_scale #Transition graphs util.plot_data_transition_gen_gamma_estimates(beta_params, alpha_params, m_params, scale_params, start, end, ages, smooth, datatype='population') util.plot_data_transition(pop_data, start, end, ages, smooth, datatype='population') util.plot_data_transition_gen_gamma_overlay_estimates(pop_data, beta_params, alpha_params, m_params, scale_params, start, end, ages, smooth, datatype='population') #Graph comparison between 2014 and 2100 util.plot_2100(beta_params, alpha_params, m_params, scale_params, ages, smooth, datatype='population')
ages,
smooth,
datatype='population')
util.plot_data_transition(pop_data,
start,
end,
ages,
smooth,
datatype='population')
util.overlay_estimates(pop_data,
a_params,
b_params,
p_params,
q_params,
scale_params,
start,
end,
ages,
smooth,
datatype='population')
#Graph comparison between 2014 and 2100
util.plot_2100(a_params,
b_params,
p_params,
q_params,
scale_params,
ages,
smooth,
datatype='population')
p_params = L_MLE_p, k_MLE_p, x_MLE_p, np.min(p_list)
#########################################
#Fit q_list to logistic function
L_0 = 0.55
k_0 = 0.31
x_0 = 1995
L_MLE_q, k_MLE_q, x_MLE_q = util.logistic_est(q_list, L_0, k_0, x_0, years, smooth, datatype='mortality', param='q', flip=True)
q_params = L_MLE_q, k_MLE_q, x_MLE_q, np.min(q_list)
#########################################
#Fit scales to logistic function
L_0 = max(scales)
k_0 = 1.5
x_0 = 1995
L_MLE_scale, k_MLE_scale, x_MLE_scale = util.logistic_est(scales, L_0, k_0, x_0, years, smooth, datatype='mortality', param='Scale')
scale_params = L_MLE_scale, k_MLE_scale, x_MLE_scale, np.min(scales)
non_infant_params = a_params, b_params, p_params, q_params, scale_params
ages = np.linspace(0, 99, 100)
#Transition graphs
util.plot_forecast_transition( (non_infant_params, infant_params), ages, start, end, smooth, datatype='mortality')
util.plot_data_transition(non_infant_mort, ages, start, end, smooth, datatype='mortality')
util.overlay_estimates(non_infant_mort, (non_infant_params, infant_params), ages, start, end, smooth, datatype='mortality')
#Graph comparison between 2014 and 2100
util.plot_2100( (non_infant_params, infant_params), ages, smooth, datatype='mortality')
pickle.dump( (non_infant_params, infant_params), open('data/demographic/parameters/mort.p', 'wb') )
f = 1.1 # Increase estimate by 10%
else:
f = 1 # Keep estimate constant
else:
f = 1 # Keep estimate constant
a = 1 / 30 * (1e-20 / (f * last_val - g) - b )
if age == -1:
b = - b / 5
age_params.append([constant, beta, a, b, c, g])
forecast = e ** (a * forecast_vals ** 2 + b * forecast_vals + c) + g
plt.plot(forecast_years, forecast, label='Forecast')
plt.plot(years, estline, label='OLS Estimate')
plt.plot(imm_age, label='Age ' + str(age))
plt.xlabel(r'Year $t$')
plt.ylabel(r'Immigration Rate $i_{s,t}$')
plt.grid(b=True, which='major', color='0.65', linestyle='-')
plt.tight_layout()
plt.legend()
plt.savefig('graphs/' + datatype + '/age_forecasts/' + str(age))
plt.close()
# Transition graphs
util.plot_forecast_transition(age_params, ages, start, end, smooth=0, datatype='immigration', options={'transition_year': 2015})
util.plot_data_transition(imm_rate, ages, start, end, smooth=0, datatype='immigration')
util.overlay_estimates(imm_rate, age_params, ages, start, end, smooth=0, datatype='immigration', options={'transition_year': 2015})
#Graph comparison between 2014 and 2030
util.plot_2100(age_params, ages, smooth=0, datatype='immigration', options={'transition_year': 2015})
pickle.dump( age_params, open('data/demographic/parameters/imm.p', 'wb') )
flip=True)
scale_params = L_MLE_scale, k_MLE_scale, x_MLE_scale, np.min(scales)
params = a_params, b_params, p_params, q_params, scale_params
#Transition graphs
util.plot_forecast_transition(params,
ages,
start,
end,
smooth,
datatype='fertility')
util.plot_data_transition(fert_data,
ages,
start,
end,
smooth,
datatype='fertility')
util.overlay_estimates(fert_data,
params,
ages,
start,
end,
smooth,
datatype='fertility')
#Graph comparison between 2014 and 2100
util.plot_2100(params, ages, smooth, datatype='fertility')
pickle.dump(params, open('data/demographic/parameters/fert.p', 'wb'))