-
Notifications
You must be signed in to change notification settings - Fork 0
/
estimation.py
132 lines (93 loc) · 3.89 KB
/
estimation.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
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
import numpy as np
import matplotlib.pyplot as plt
import numba as nb
import pickle
import pandas as pd
from numpy import exp, log
from math import erfc, sqrt
from numba import jit, njit, typed
from scipy.optimize import minimize
from basic import std_norm_cdf, tauchen, \
interp2d_lin, interp2d_exp, interp3d_lin, interp3d_exp, interp_lin, interp_exp, \
getavggrid, wtd_avg_and_std, wtd_cov, weighted_quantile
from fcn import getrhovec, \
u, du, \
Scf, dScf, Ssf, dSsf, f, fc, fs, \
rapf, rpf, ramf, rmf
from VFI import initdEVs, getgrids, VFIsolve
from simulation import gettransmats, genergdist, \
getmom1, getcldist, getmom2, getmom3, \
genapqdist, gennegnwdist, gennegnwdistpb,\
getmom4, getmom5, getmom6, \
getallmoms
extpar, polpar, _, gridpar = pickle.load(open('../vars/params.pickle', 'rb'))
init_dEVsgrid = pickle.load(open('../vars/init_dEVsgrid.pickle', 'rb'))
lb = np.array([1., .2, 0., 1., 1., 1., 0., -2., 0., 0., 0., 1e-4, 0., 0.5, -0.1])
ub = np.array([2.5, 2., 1., 2., 3., 10., 1., 1., 1., 1.1, 3., 1e-1, 1., 1., 0.2])
lb_hs = np.array([-2., 0., -1])
ub_hs = np.array([0., 10., 1])
def estpar_to_estparhat(par):
parhat = log(1/(1-(par-lb)/(ub-lb)) - 1)
return parhat
def estparhat_to_estpar(parhat):
par = (ub-lb)*(1-1/(exp(parhat)+1))+lb
return par
def smalllossf(estparhat_hs, cldistpc, negnwdist, simpc, dataM):
hs_coeffs = (ub_hs-lb_hs)*(1-1/(exp(estparhat_hs)+1))+lb_hs
smallhatM = getmom5(cldistpc, negnwdist, simpc, hs_coeffs)
smalldataM = dataM[15:19]
I = np.eye(len(smalldataM))
err = ((smallhatM - smalldataM) @ I) @ (smallhatM - smalldataM)
return err
def lossf(parhat, dataM, init_hs_coeffs, config):
par = estparhat_to_estpar(parhat)
Peta, Peps, etagrid, epsgrid, \
icgrid, isgrid, hgrid,\
asgrid, apgrid, \
zsgrid, \
zphatgrid, zphat2grid,\
zmgrid, zmhatgrid = getgrids(par, polpar)
dEVsgrid, \
solveVmgrid,\
solvetilVpgrid,\
pngrid,\
solvetilVsgrid,\
colgrid = VFIsolve(par, polpar, init_dEVsgrid)
transabmat, \
transpmat, sims,\
tranpabmat, tranpbcmat, tranpacmat,\
tranpmmat, simpb, simpc, \
tranmsmat, simm, \
tranallmat = gettransmats(solveVmgrid,\
solvetilVpgrid,\
pngrid,\
solvetilVsgrid,\
colgrid,\
par, polpar)
erglam, ergdistall, ergdistsa, ergdistsb, ergdistpa, ergdistpb, ergdistpc, ergdistmb \
= genergdist(tranallmat, transabmat, transpmat, tranpabmat, tranpbcmat, tranpmmat, tranmsmat)
cldistsb, cldistpa, cldistpb, cldistpc, cldistm, mcldistsa, mcldistsb, cmcldistsb \
= getcldist(ergdistsb, sims, transpmat, tranpabmat, tranpacmat, tranpmmat, tranmsmat, transabmat)
negnwdist, mnegnwdist = gennegnwdist(ergdistsb, sims, transpmat, tranpacmat, tranpmmat, tranmsmat, transabmat)
estparhat_hs = log(1/(1-(init_hs_coeffs-lb_hs)/(ub_hs-lb_hs)) - 1)
output = minimize(smalllossf, estparhat_hs, args = (cldistpc, negnwdist, simpc, dataM), method='Nelder-Mead')
hs_coeffs = (ub_hs-lb_hs)*(1-1/(exp(output.x)+1))+lb_hs
hatM = getallmoms(ergdistsb, ergdistpc, \
cldistsb, cldistpc, mcldistsb, cmcldistsb, \
negnwdist, \
sims, simpc, \
etagrid, par, polpar, hs_coeffs)
if config == 0:
tmp = np.ones(len(hatM))
elif config == 1:
tmp = np.ones(len(hatM))
tmp[5:11] = 1e-1
elif config == 2:
tmp = 1e-1*np.ones(len(hatM))
tmp[5:11] = 1
W = np.diag(tmp)
tmp = dataM
err = ((hatM - tmp) @ W) @ (hatM - tmp)
print(err)
print(par)
return err, hs_coeffs