-
Notifications
You must be signed in to change notification settings - Fork 0
/
synth_negbinom.py
executable file
·102 lines (69 loc) · 3 KB
/
synth_negbinom.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
import numpy as np
from scipy.optimize import minimize
from tools import (meta_population_sample, meta_population_cumulative,
load_full_dataset, load_specific_city, data_travel)
from config import get_arguments
# Params
args = get_arguments()
start_date = args.start
end_date = args.end
active_sampling = args.active_sampling
t_max = 2300
M_c = args.c
cities = ['Tartagal', 'Oran', 'Jujuy', 'Guemes', 'Tucuman', 'Santa Fe', 'Mendoza', 'Buenos Aires']
cities = cities[:M_c]
Ns = np.array([4.4e4, 5.1e4, 2e5, 2.3e4, 6.3e5, 4e5, 8e5, 2e6])
Ns = Ns[:M_c]
travel_matrix = data_travel() / args.travel_norm
travel_matrix = travel_matrix[:M_c, :M_c]
params = [args.asym,
np.array([args.betas2[0], args.betas2[0], args.betas2[1]]),
args.gamma, args.phi, args.sin_0, args.init, args.shift,
Ns, M_c, travel_matrix]
# Functions
def objective(pars, # asym, beta_1, beta_2, t_0, init, phi
shift, gamma, Ns, M_c, travel_matrix, sampling, y_sample, opt_ids):
betas = np.array([pars[1], pars[1], pars[2]])
y_opt = meta_population_sample(sampling, pars[0], betas, gamma, pars[5], pars[3], pars[4], shift,
Ns, M_c, travel_matrix, init_n=M_c, active_sampling=active_sampling)
if args.r > 1:
y_opt = np.repeat(y_opt, args.r)
objs = (y_opt - y_sample)[opt_ids] ** 2.0
return objs.sum() / objs.shape[0]
# Load Sample
data = load_full_dataset()
sampling = dict()
y_sample = []
for j, city in enumerate(cities):
x, z = load_specific_city(data, city, start_date, end_date)
sampling[city] = x[z > 0.0]
y_sample.append(meta_population_cumulative(x[z > 0.0], *params,
init_n=M_c, active_sampling=active_sampling)[:, 4 + 5 * j])
y_sample = np.concatenate(y_sample)
y_sample = np.repeat(y_sample, args.r)
pp = y_sample / args.noise**2.0
nn = pp * y_sample / (1.0 - pp)
y_sample = np.random.negative_binomial(nn, pp)
if not args.negative_cases:
y_sample[y_sample < 0.0] = 0.0
if args.neglect_zeros:
opt_ids = y_sample > 0.0
else:
opt_ids = np.repeat(True, y_sample.shape[0])
# Minimization
# x0 = np.array([args.asym, args.beta, args.sin_0, args.init, args.phi, args.sigma])
arguments = (args.shift, args.gamma, Ns, M_c, travel_matrix, sampling, y_sample, opt_ids)
limits = [[0.0, 1.0], [0.0, None], [0.0, None], [0.0, None], [0.0, None], [0.0, 2.0 * np.pi]]
ws = []
params = [args.asym, *args.betas2, args.sin_0, args.init, args.phi, args.noise, args.shift, args.gamma]
for ii in range(args.m):
x0 = np.array([np.random.random(), np.random.random() * 0.179 + 0.001, np.random.random() * 0.179 + 0.001,
np.random.random() * 20.0, np.random.random() * 20.0, np.random.random() * 2.0 * np.pi])
res = minimize(objective, x0, args=arguments, bounds=limits)
temp_sd = objective(res.x, *arguments)
ws.append(list(res.x))
ws[-1].append(temp_sd)
ws[-1] = params + ws[-1]
ws = np.array(ws)
# Save
np.save('results/' + args.name + '.npy', ws)