-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathhdtest.py
320 lines (262 loc) · 6.6 KB
/
hdtest.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
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
#%%
import numpy
import scipy
import os
os.chdir('/home/bbales2/diffusion')
import pyximport
import time
pyximport.install(reload_support = True)
import matplotlib.pyplot as plt
data = """0 2.496643737
5.6568 2.760147955
12.0599 2.731482238
18.4631 2.59114926
24.1199 2.48375488
29.7768 2.432533019
36.1799 2.240920246
41.8368 2.232590513
48.2399 2.161828643
53.8967 2.059430461
60.2999 1.996965007
65.9567 1.887728518
72.3598 1.907844208
78.017 1.808829964
84.42 1.57142775
90.077 1.701358373
95.734 1.605451805
102.137 1.456515395
108.54 1.461104922
114.197 1.352972172
119.854 1.302182066
126.257 1.243359915
132.66 1.066671415
138.317 1.159614932
143.973 1.126754173
151.045 1.070932217
156.701 1.032455722
162.358 1.045294806
168.015 0.96000336
174.418 0.893994653
180.821 0.903897486
186.478 0.830675659
192.135 0.774701278
198.538 0.718758126
204.941 0.716032941
210.598 0.713599868
216.255 0.630989886
222.658 0.648870892
228.315 0.567879499
234.718 0.541446439
240.375 0.548667775
246.778 0.468353856
252.435 0.495793694
258.838 0.450233883
264.495 0.435563393
270.898 0.403205311
276.555 0.427295594
282.212 0.359591388
289.283 -0.582969908
294.94 0.335402222
300.597 0.309952966
306.254 0.300129552
313.325 0.293129426
318.981 0.293583798
324.638 0.239513564
330.295 0.255303141
336.698 0.220204387
343.101 0.225505408
348.758 0.227246057
354.415 0.154325388
360.818 0.199752451
367.221 0.169280343
372.878 0.176321289
378.535 0.167477769
384.938 0.145899825
390.595 0.125536605
396.998 0.12135882
402.655 0.114396878
409.058 0.11778856
414.715 0.118246595
421.118 0.102681041
426.775 0.085488536
433.178 0.070239495
438.835 0.093348714
445.238 0.088575438
451.641 0.077767637
457.298 0.078043344
462.955 0.071393113
468.612 0.041080937
475.683 0.058835859
481.34 0.048843231
486.997 0.051745413
492.654 0.06697217"""
dataList = []
for line in data.split('\n')[0::2]:
d, at = line.split()
d = float(d)
at = float(at)
if at > 0.0:
dataList.append((d, at))
data = numpy.array(dataList)
xs = numpy.array(data[1:, 0])
at = numpy.array(data[1:, 1])
nxs = numpy.linspace(0.0, max(xs), len(xs))
at = numpy.interp(nxs, xs, at)
xs = nxs
maxxs = max(xs)
maxat = max(at)
xs /= max(xs)
at /= max(at)
#%%
N = len(xs)
#xs = numpy.linspace(0, 1, N + 2)
dx = xs[1] - xs[0]
D = None
a = 1.0
b1 = 1.0
b2 = 0.0
#for a in [-0.0001, 0.0, 0.0001]:
def f(y, t0, D, a):
dcdt = numpy.zeros(N)
co = y
f = (D * (co[1] - 2 * co[0] + b1) / dx**2 + D * (a / 2.0) * (co[1]**2 - 2 * co[0]**2 + b1**2) / dx**2)
dcdt[0] = f
for i in range(1, N - 1):
f = (D * (co[i + 1] - 2 * co[i] + co[i - 1]) / dx**2 + D * (a / 2.0) * (co[i + 1]**2 - 2 * co[i]**2 + co[i - 1]**2) / dx**2)
dcdt[i] = f
f = (D * (b2 - 2 * co[N - 1] + co[N - 2]) / dx**2 + D * (a / 2.0) * (b2**2 - 2 * co[N - 1]**2 + co[N - 2]**2) / dx**2)
dcdt[N - 1] = f
return dcdt
def solve(D, a):
return scipy.integrate.odeint(f, numpy.zeros(N), [0.0, 0.1], args = (D, a), mxstep = 1000)[1]
plt.plot(solve(1.0, 0.2))
plt.show()
#%%
zs = numpy.linspace(0, 5.0, 100)
Dalpha = 0.4
Dbeta = 1.4
aalpha = -0.5
abeta = 1.0
D = lambda x : (Dbeta - Dalpha) * (x + 1.0) / 2.0 + Dalpha
a = lambda y : (abeta - aalpha) * (y + 1.0) / 2.0 + aalpha
zD = lambda D : 2.0 * (D - Dalpha) / (Dbeta - Dalpha) - 1.0
za = lambda a : 2.0 * (a - aalpha) / (abeta - aalpha) - 1.0
#%%
os.chdir('/home/bbales2/gpc')
import gpc
reload(gpc)
minD = 0.4
maxD = 1.4
mina = -0.5
maxa = 1.0
hd = gpc.GPC(7, solve, [('u', (minD, maxD), 7),
('u', (mina, maxa), 7)])
#%%
mn = 0.2
def L(z1, z2):
logp = 0.0
u = hd.approx(z1, z2)
for i in range(1, len(u)):
logp += -0.5 * (at[i] - u[i])**2 / mn**2 - numpy.log(mn) - 0.5 * numpy.log(2.0) - 0.5 * numpy.log(numpy.pi)
return numpy.exp(logp)
denominator = hd.measure(L)
post = numpy.zeros((30, 30))
for i, z1 in enumerate(numpy.linspace(minD, maxD, 30)):
for j, z2 in enumerate(numpy.linspace(mina, maxa, 30)):
post[i, j] = L(z1, z2) * hd.prior(z1, z2) / denominator#
plt.imshow(post, interpolation = 'NONE', extent = [mina, maxa, maxD, minD])
plt.colorbar()
plt.xlabel('a')
plt.ylabel('D')
plt.title('Posterior')
plt.gcf().set_size_inches((12, 8))
plt.show()
#%%
plt.plot(hd.approx(0.69, 0.3));
plt.plot(at);#solve(0.89, 0.3));
#%%
def loss(z):
z1 = z[0]
z2 = z[1]
print z1, z2
return -L(z1, z2) * hd.prior(z1, z2)
res = scipy.optimize.minimize(loss, [0.8, -0.3], bounds = ([minD, maxD - 0.01], [mina, maxa - 0.01]))#
#%%
plt.plot(xs, at, 'bo')
plt.plot(xs, solve(0.89178849, -0.3399514), 'r')
plt.plot(xs, solve(0.88, 0.0), 'g')
plt.legend(['data', 'opt', 'maxp'])
plt.show()
#%%
os.chdir('/home/bbales2/gpc')
import gpc
reload(gpc)
minD = 0.5
maxD = 1.2
mina = -0.5
maxa = 0.5
solve2 = lambda a : solve(1.0, a)
hd = gpc.GPC(3, solve2, [('n', (-0.2, 0.2), 3)])
plt.plot(hd.approx(-0.2), 'r')
plt.plot(hd.sols[(0.0,)], 'g')
plt.plot(solve2(-0.2), 'b')
plt.show()
#%%
labels = []
for z1 in numpy.linspace(minD, maxD, 5):
z2 = mina
plt.plot(hd.approx(z1, z2), 'r')
plt.plot(solve(z1, z2), 'b--')
labels.append("D = {0}, a = {1}".format(z1, z2))
plt.legend(labels)
plt.show()
labels = []
for z2 in numpy.linspace(mina, maxa, 5):
z1 = maxD
plt.plot(hd.approx(z1, z2), 'r')
plt.plot(solve(z1, z2), 'b--')
labels.append("D = {0}, a = {1}".format(z1, z2))
plt.legend(labels)
plt.show()
#%%
plt.plot(approx(zD(0.85), za(-0.05)), 'r')
plt.plot(approx(zD(0.89), za(-0.34)), 'g')
plt.plot(at, 'bo')
plt.plot(solve(0.89, -0.34), 'rx')
plt.show()
#%%
reload(gpc)
import scipy.stats
#deg = 10
#xh, wh = scipy.special.p_roots(deg)
for x, wx in zip(xxh, wxh):
for y, wy in zip(xyh, wyh):
denominator += wx * wy * L(x, y) / ((Dbeta - Dalpha) * (abeta - aalpha))
#plt.plot(approx(x))
#plt.plot(data, '*')
#plt.title("D = {0}, z = {1}".format(D(x), x))
#plt.show()
denominator /= (1.0 / ((Dbeta - Dalpha) * (abeta - aalpha)))
zs = numpy.linspace(-1.0, 1.0, 100)
post = numpy.zeros((len(zs), len(zs)))
for i, z1 in enumerate(zs):
for j, z2 in enumerate(zs):
post[i, j] = L(z1, z2) * (1.0 / ((Dbeta - Dalpha) * (abeta - aalpha))) / denominator
plt.imshow(post, interpolation = 'NONE', extent = [aalpha, abeta, Dbeta, Dalpha])
plt.colorbar()
plt.xlabel('a')
plt.ylabel('D')
plt.title('Posterior')
plt.gcf().set_size_inches((12, 8))
#plt.plot(D(zs), [(1.0 / ((Dbeta - Dalpha) * (abeta - aalpha)))] * len(zs), 'b')
#plt.legend(['Posterior', 'Prior'])
plt.show()
#%%
total = 0.0
for x, w in zip(xh, wh):
total += w * L(x) * (1.0 / (beta - alpha)) / denominator
print total / (1.0 / (beta - alpha))
#%%
#%%
plt.plot(solve(mu)[0])
plt.show()