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
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
|
"""
Test series for lowess, stl and loess routines.
:author: Pierre GF Gerard-Marchant
:contact: pierregm_at_uga_edu
:date: $Date: 2007-03-26 23:38:36 -0700 (Mon, 26 Mar 2007) $
:version: $Id: test_mpyloess.py 2874 2007-03-27 06:38:36Z pierregm $
"""
__author__ = "Pierre GF Gerard-Marchant ($Author: pierregm $)"
__version__ = '1.0'
__revision__ = "$Revision: 2874 $"
__date__ = '$Date: 2007-03-26 23:38:36 -0700 (Mon, 26 Mar 2007) $'
import os
import numpy
from numpy import bool_, complex_, float_, int_, str_, object_
import numpy.core.numeric as numeric
fromiter = numpy.fromiter
import maskedarray
marray = maskedarray.masked_array
masked_values = maskedarray.masked_values
from numpy.testing import NumpyTest, NumpyTestCase
from maskedarray.testutils import build_err_msg, \
assert_equal, assert_almost_equal
import mpyloess
reload(mpyloess)
from mpyloess import lowess, stl, loess, loess_anova
#####---------------------------------------------------------------------------
#---- --- LOWESS ---
#####---------------------------------------------------------------------------
class test_lowess(NumpyTestCase):
"Test class for lowess."
#
def __init__(self, *args, **kwds):
NumpyTestCase.__init__(self, *args, **kwds)
X = marray([ 1, 2, 3, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 8,10,12,14,50])
Y = marray([18, 2,15, 6,10, 4,16,11, 7, 3,14,17,20,12, 9,13, 1, 8, 5,19])
idx = X.argsort()
self.data = (X[idx], Y[idx])
#............................................
def test_lowess_1(self):
"Tests lowess on typical data. part #1."
(X, Y) = self.data
YS = [13.659,11.145, 8.701, 9.722,10.000,11.300,11.300,11.300,
11.300,11.300,11.300,11.300,11.300,11.300,11.300,13.000,
6.440, 5.596, 5.456,18.998]
Z = lowess(X, Y, span=0.25, nsteps=0, delta=0)
assert_almost_equal(Z.outputs.fitted_values, YS, decimal=3)
assert_almost_equal(Z.outputs.fitted_residuals+Z.outputs.fitted_values,
Z.inputs.y, decimal=3)
#............................................
def test_lowess_2(self):
"Tests lowess on typical data. part #2."
(X, Y) = self.data
YS = [13.659,12.347,11.034, 9.722,10.511,11.300,11.300,11.300,
11.300,11.300,11.300,11.300,11.300,11.300,11.300,13.000,
6.440, 5.596, 5.456,18.998]
Z = lowess(X, Y, span=0.25, nsteps=0, delta=3)
assert_almost_equal(Z.outputs.fitted_values, YS, decimal=3)
assert_almost_equal(Z.outputs.fitted_residuals+Z.outputs.fitted_values,
Z.inputs.y, decimal=3)
#............................................
def test_lowess_3(self):
"Tests lowess on typical data. part #3."
(X, Y) = self.data
YS = [14.811,12.115, 8.984, 9.676,10.000,11.346,11.346,11.346,
11.346,11.346,11.346,11.346,11.346,11.346,11.346,13.000,
6.734, 5.744, 5.415,18.998 ]
Z = lowess(X, Y, span=0.25, nsteps=2, delta=0)
assert_almost_equal(Z.outputs.fitted_values, YS, decimal=3)
assert_almost_equal(Z.outputs.fitted_residuals+Z.outputs.fitted_values,
Z.inputs.y, decimal=3)
#............................................
def test_lowess_4(self):
"Tests lowess on masked data."
X = masked_values([ 1, 2, 3, 4, 5,-999, 6, 6, 6, 6, -999,-999,
6, 6, 6, 6, 6, 6, 8,-999,10,12,14,50,-999],-999)
Y = marray([18, 2,15, 6,10,-999, 4, 16,11, 7, -999,-999,
3,14,17,20,12, 9,13,-999, 1, 8, 5,19,-999])
YS = [14.811,12.115, 8.984, 9.676,10.000,11.346,11.346,11.346,
11.346,11.346,11.346,11.346,11.346,11.346,11.346,13.000,
6.734, 5.744, 5.415,18.998 ]
Z = lowess(X, Y, span=0.25, nsteps=2, delta=0)
assert_almost_equal(Z.outputs.fitted_values.compressed(), YS, decimal=3)
assert_almost_equal(Z.outputs.fitted_residuals + Z.outputs.fitted_values,
Z.inputs.y, decimal=3)
#
#####---------------------------------------------------------------------------
#---- --- STL ---
#####---------------------------------------------------------------------------
class test_stl(NumpyTestCase):
"Tests STL."
#
def __init__(self, *args, **kwds):
NumpyTestCase.__init__(self, *args, **kwds)
# Get CO2 data ................
filename = os.path.join('tests','co2_data')
F = open(filename, 'r')
data = []
for line in F.readlines():
data.append([float(x) for x in line.rstrip().split()])
co2_data = numpy.concatenate(data)
# Get CO2 results .............
filename = os.path.join('tests','co2_results_double')
F = open(filename, 'r')
co2_results = []
for line in F.readlines():
co2_results.append(fromiter((float(x) for x in line.rstrip().split()),
float_))
#
parameters = dict(np=12, ns=35, nt=19, nl=13, no=2, ni=1,
nsjump=4, ntjump=2, nljump=2,
isdeg=1, itdeg=1, ildeg=1)
self.d = (co2_data, co2_results, parameters)
#............................................
def test_stl_1(self):
"Tests a classic STL."
(co2_data, co2_results, parameters) = self.d
co2_fitted = stl(co2_data, robust=False, **parameters).outputs
assert_almost_equal(co2_fitted.seasonal, co2_results[0], 6)
assert_almost_equal(co2_fitted.trend, co2_results[1], 6)
assert_almost_equal(co2_fitted.weights, co2_results[2], 6)
#............................................
def test_stl_2(self):
"Tests a robust STL."
(co2_data, co2_results, parameters) = self.d
co2_fitted = stl(co2_data, robust=True, **parameters).outputs
assert_almost_equal(co2_fitted.seasonal, co2_results[4], 6)
assert_almost_equal(co2_fitted.trend, co2_results[5], 6)
assert_almost_equal(co2_fitted.weights, co2_results[6], 6)
#####---------------------------------------------------------------------------
#---- --- LOESS ---
#####---------------------------------------------------------------------------
class test_loess2d(NumpyTestCase):
"Test class for lowess."
#
def __init__(self, *args, **kwds):
NumpyTestCase.__init__(self, *args, **kwds)
dfile = open(os.path.join('tests','madeup_data'), 'r')
dfile.readline()
x = fromiter((float(v) for v in dfile.readline().rstrip().split()),
float_).reshape(-1,2)
x = marray(x)
dfile.readline()
y = fromiter((float(v) for v in dfile.readline().rstrip().split()),
float_)
y = marray(y)
#
rfile = open(os.path.join('tests','madeup_result'), 'r')
results = []
for i in range(8):
rfile.readline()
z = fromiter((float(v) for v in rfile.readline().rstrip().split()),
float_)
results.append(z)
#
newdata1 = numpy.array([[-2.5, 0.0, 2.5], [0., 0., 0.]])
newdata2 = numpy.array([[-0.5, 0.5], [0., 0.]])
#
madeup = loess(x,y)
self.d = (x, y, results, newdata1, newdata2, madeup)
#
def test_2dbasic(self):
"2D standard"
(x, y, results, _, _, madeup) = self.d
madeup = loess(x,y)
madeup.model.span = 0.5
madeup.model.normalize = True
madeup.fit()
assert_almost_equal(madeup.outputs.fitted_values, results[0], 5)
assert_almost_equal(madeup.outputs.enp, 14.9, 1)
assert_almost_equal(madeup.outputs.s, 0.9693, 4)
#
# def test_2d_modflags_ez(self):
# "2D - modification of model flags"
# (x, y, results, newdata1, newdata2, madeup) = self.d
# madeup = cloess.loess(x,y)
# madeup.model.span = 0.8
# madeup.model.drop_square_flags[0] = True
# madeup.model.parametric_flags[0] = True
# assert_equal(madeup.model.parametric_flags[:2],[1,0])
# madeup.fit()
# assert_almost_equal(madeup.outputs.fitted_values, results[1], 5)
# assert_almost_equal(madeup.outputs.enp, 6.9, 1)
# assert_almost_equal(madeup.outputs.s, 1.4804, 4)
#
def test_2d_modflags_tot(self):
"2D - modification of model flags"
(x, y, results, _, _, madeup) = self.d
madeup = loess(x,y)
madeup.model.span = 0.8
madeup.model.drop_square_flags = [True, False]
madeup.model.parametric_flags = [True, False]
assert_equal(madeup.model.parametric_flags[:2],[1,0])
madeup.fit()
assert_almost_equal(madeup.outputs.fitted_values, results[1], 5)
assert_almost_equal(madeup.outputs.enp, 6.9, 1)
assert_almost_equal(madeup.outputs.s, 1.4804, 4)
#
def test_2d_modfamily(self):
"2D - family modification"
(_, _, results, _, _, madeup) = self.d
madeup.model.span = 0.8
madeup.model.drop_square_flags = [True, False]
madeup.model.parametric_flags = [True, False]
madeup.model.family = "symmetric"
madeup.fit()
assert_almost_equal(madeup.outputs.fitted_values, results[2], 5)
assert_almost_equal(madeup.outputs.enp, 6.9, 1)
assert_almost_equal(madeup.outputs.s, 1.0868, 4)
#
def test_2d_modnormalize(self):
"2D - normalization modification"
(_, _, results, _, _, madeup) = self.d
madeup.model.span = 0.8
madeup.model.drop_square_flags = [True, False]
madeup.model.parametric_flags = [True, False]
madeup.model.family = "symmetric"
madeup.model.normalize = False
madeup.fit()
assert_almost_equal(madeup.outputs.fitted_values, results[3], 5)
assert_almost_equal(madeup.outputs.enp, 6.9, 1)
assert_almost_equal(madeup.outputs.s, 1.0868, 4)
#
def test_2d_pred_nostderr(self):
"2D prediction - no stderr"
(_, _, results, newdata1, _, madeup) = self.d
madeup.model.span = 0.5
madeup.model.normalize = True
madeup.predict(newdata1, stderror=False)
assert_almost_equal(madeup.predicted.values, results[4], 5)
#
madeup_pred = madeup.predict(newdata1, stderror=False)
assert_almost_equal(madeup_pred.values, results[4], 5)
#
def test_2d_pred_nodata(self):
"2D prediction - nodata"
(_, _, _, _, _, madeup) = self.d
try:
madeup.predict(None)
except ValueError:
pass
else:
raise AssertionError,"The test should have failed"
#
def test_2d_pred_stderr(self):
"2D prediction - w/ stderr"
(_, _, results, _, newdata2, madeup) = self.d
madeup.model.span = 0.5
madeup.model.normalize = True
madeup_pred = madeup.predict(newdata2, stderror=True)
assert_almost_equal(madeup_pred.values, results[5], 5)
assert_almost_equal(madeup_pred.stderr, [0.276746, 0.278009], 5)
assert_almost_equal(madeup_pred.residual_scale, 0.969302, 6)
assert_almost_equal(madeup_pred.df, 81.2319, 4)
# Direct access
madeup.predict(newdata2, stderror=True)
assert_almost_equal(madeup.predicted.values, results[5], 5)
assert_almost_equal(madeup.predicted.stderr, [0.276746, 0.278009], 5)
assert_almost_equal(madeup.predicted.residual_scale, 0.969302, 6)
assert_almost_equal(madeup.predicted.df, 81.2319, 4)
#
def test_2d_pred_confinv(self):
"2D prediction - confidence"
(_, _, results, _, newdata2, madeup) = self.d
madeup.model.span = 0.5
madeup.model.normalize = True
madeup_pred = madeup.predict(newdata2, stderror=True)
madeup.predicted.confidence(coverage=0.99)
assert_almost_equal(madeup.predicted.confidence_intervals.lower,
results[6][::3], 5)
assert_almost_equal(madeup.predicted.confidence_intervals.fit,
results[6][1::3], 5)
assert_almost_equal(madeup.predicted.confidence_intervals.upper,
results[6][2::3], 5)
# Direct access
confinv = madeup.predicted.confidence(coverage=0.99)
assert_almost_equal(confinv.lower, results[6][::3], 5)
assert_almost_equal(confinv.fit, results[6][1::3], 5)
assert_almost_equal(confinv.upper, results[6][2::3], 5)
#####---------------------------------------------------------------------------
#---- --- test 1D ---
#####---------------------------------------------------------------------------
class test_loess_gas(NumpyTestCase):
"Test class for lowess."
#
def __init__(self, *args, **kwds):
NumpyTestCase.__init__(self, *args, **kwds)
NOx = marray([4.818, 2.849, 3.275, 4.691, 4.255, 5.064, 2.118, 4.602,
2.286, 0.970, 3.965, 5.344, 3.834, 1.990, 5.199, 5.283,
3.752, 0.537, 1.640, 5.055, 4.937, 1.561])
E = marray([0.831, 1.045, 1.021, 0.970, 0.825, 0.891, 0.71, 0.801,
1.074, 1.148, 1.000, 0.928, 0.767, 0.701, 0.807, 0.902,
0.997, 1.224, 1.089, 0.973, 0.980, 0.665])
gas_fit_E = numpy.array([0.665, 0.949, 1.224])
newdata = numpy.array([0.6650000, 0.7581667, 0.8513333, 0.9445000,
1.0376667, 1.1308333, 1.2240000])
coverage = 0.99
rfile = open(os.path.join('tests','gas_result'), 'r')
results = []
for i in range(8):
rfile.readline()
z = fromiter((float(v) for v in rfile.readline().rstrip().split()),
float_)
results.append(z)
self.d = (E, NOx, gas_fit_E, newdata, coverage, results)
#
def test_1dbasic(self):
"Basic test 1d"
(E, NOx, _, _, _, results) = self.d
gas = loess(E,NOx)
gas.model.span = 2./3.
gas.fit()
assert_almost_equal(gas.outputs.fitted_values, results[0], 6)
assert_almost_equal(gas.outputs.enp, 5.5, 1)
assert_almost_equal(gas.outputs.s, 0.3404, 4)
#
def test_1dbasic_alt(self):
"Basic test 1d - part #2"
(E, NOx, _, _, _, results) = self.d
gas_null = loess(E, NOx)
gas_null.model.span = 1.0
gas_null.fit()
assert_almost_equal(gas_null.outputs.fitted_values, results[1], 5)
assert_almost_equal(gas_null.outputs.enp, 3.5, 1)
assert_almost_equal(gas_null.outputs.s, 0.5197, 4)
#
def test_1dpredict(self):
"Basic test 1d - prediction"
(E, NOx, gas_fit_E, _, _, results) = self.d
gas = loess(E,NOx, span=2./3.)
gas.fit()
gas.predict(gas_fit_E, stderror=False)
assert_almost_equal(gas.predicted.values, results[2], 6)
#
def test_1dpredict_2(self):
"Basic test 1d - new predictions"
(E, NOx, _, newdata, _, results) = self.d
gas = loess(E,NOx, span=2./3.)
gas.predict(newdata, stderror=True)
gas.predicted.confidence(0.99)
assert_almost_equal(gas.predicted.confidence_intervals.lower,
results[3][0::3], 6)
assert_almost_equal(gas.predicted.confidence_intervals.fit,
results[3][1::3], 6)
assert_almost_equal(gas.predicted.confidence_intervals.upper,
results[3][2::3], 6)
#
def test_anova(self):
"Tests anova"
(E, NOx, _, _, _, results) = self.d
gas = loess(E,NOx, span=2./3.)
gas.fit()
gas_null = loess(E, NOx, span=1.0)
gas_null.fit()
gas_anova = loess_anova(gas, gas_null)
gas_anova_theo = results[4]
assert_almost_equal(gas_anova.dfn, gas_anova_theo[0], 5)
assert_almost_equal(gas_anova.dfd, gas_anova_theo[1], 5)
assert_almost_equal(gas_anova.F_value, gas_anova_theo[2], 5)
assert_almost_equal(gas_anova.Pr_F, gas_anova_theo[3], 5)
#
def test_failures(self):
"Tests failures"
(E, NOx, gas_fit_E, _, _, _) = self.d
gas = loess(E,NOx, span=2./3.)
# This one should fail (all parametric)
gas.model.parametric_flags = True
self.assertRaises(ValueError, gas.fit)
# This one also (all drop_square)
gas.model.drop_square_flags = True
self.assertRaises(ValueError, gas.fit)
gas.model.degree = 1
self.assertRaises(ValueError, gas.fit)
# This one should not (revert to std)
gas.model.parametric_flags = False
gas.model.drop_square_flags = False
gas.model.degree = 2
gas.fit()
# Now, for predict .................
gas.predict(gas_fit_E, stderror=False)
# This one should fail (extrapolation & blending)
self.assertRaises(ValueError,
gas.predict, gas.predicted.values, stderror=False)
# But this one should not ..........
gas.predict(gas_fit_E, stderror=False)
#
def test_mask(self):
NOx = marray([4.818, 2.849, 3.275, 4.691, 4.255, 5.064, 2.118, 4.602,
2.286, 0.970, 3.965, 5.344, 3.834, 1.990, 5.199, 5.283,
-9999, -9999, 3.752, 0.537, 1.640, 5.055, 4.937, 1.561])
NOx = maskedarray.masked_values(NOx, -9999)
E = marray([0.831, 1.045, 1.021, 0.970, 0.825, 0.891, 0.71, 0.801,
1.074, 1.148, 1.000, 0.928, 0.767, 0.701, 0.807, 0.902,
-9999, -9999, 0.997, 1.224, 1.089, 0.973, 0.980, 0.665])
gas_fit_E = numpy.array([0.665, 0.949, 1.224])
newdata = numpy.array([0.6650000, 0.7581667, 0.8513333, 0.9445000,
1.0376667, 1.1308333, 1.2240000])
coverage = 0.99
rfile = open(os.path.join('tests','gas_result'), 'r')
results = []
for i in range(8):
rfile.readline()
z = fromiter((float(v) for v in rfile.readline().rstrip().split()),
float_)
results.append(z)
#
gas = loess(E,NOx)
gas.model.span = 2./3.
gas.fit()
assert_almost_equal(gas.outputs.fitted_values.compressed(), results[0], 6)
assert_almost_equal(gas.outputs.enp, 5.5, 1)
assert_almost_equal(gas.outputs.s, 0.3404, 4)
########################################################################
if __name__ == '__main__':
NumpyTest().run()
|
