-
Notifications
You must be signed in to change notification settings - Fork 0
/
util.py
438 lines (395 loc) · 18.5 KB
/
util.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
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
##
#
##
import os
import matplotlib
matplotlib.use("Agg")
matplotlib.rcParams['agg.path.chunksize'] = 10000
import datetime as dt
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from scipy.interpolate import interp1d
import math
from scipy.stats import pearsonr
from sklearn.preprocessing import MinMaxScaler
from scipy.stats import norm
np.random.seed(7)
import database as db
from sklearn.dummy import DummyClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import LinearSVC
from sklearn.tree import DecisionTreeClassifier, ExtraTreeClassifier
from sklearn.neighbors import KNeighborsClassifier, RadiusNeighborsClassifier, NearestCentroid
from sklearn.ensemble import AdaBoostClassifier, BaggingClassifier, ExtraTreesClassifier, GradientBoostingClassifier, RandomForestClassifier
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis, QuadraticDiscriminantAnalysis
from sklearn.neural_network import MLPClassifier
from sklearn.model_selection import train_test_split
from sklearn.metrics import roc_curve, auc, confusion_matrix
from sklearn.dummy import DummyRegressor
from sklearn.linear_model import LinearRegression, ElasticNet, BayesianRidge
from sklearn.svm import LinearSVR
from sklearn.tree import DecisionTreeRegressor, ExtraTreeRegressor
from sklearn.neighbors import KNeighborsRegressor
from sklearn.ensemble import AdaBoostRegressor, BaggingRegressor, ExtraTreesRegressor, GradientBoostingRegressor, RandomForestRegressor
from sklearn.gaussian_process import GaussianProcessRegressor
from sklearn.gaussian_process.kernels import RBF, Matern, RationalQuadratic as RQ
from sklearn.externals import joblib
from spacepy import plot as splot
from imblearn.under_sampling import RandomUnderSampler
import verify
from verify import Contingency2x2
from keras.models import Sequential
from keras.layers import Dense,LSTM,Embedding,Dropout
def nan_helper(y):
nans = np.isnan(y)
x = lambda z: z.nonzero()[0]
f = interp1d(x(~nans), y[~nans], kind="cubic")
y[nans] = f(x(nans))
return y
def smooth(x,window_len=51,window="hanning"):
if x.ndim != 1: raise ValueError, "smooth only accepts 1 dimension arrays."
if x.size < window_len: raise ValueError, "Input vector needs to be bigger than window size."
if window_len<3: return x
if not window in ["flat", "hanning", "hamming", "bartlett", "blackman"]: raise ValueError, "Window is on of 'flat', 'hanning', 'hamming', 'bartlett', 'blackman'"
s = np.r_[x[window_len-1:0:-1],x,x[-2:-window_len-1:-1]]
if window == "flat": w = numpy.ones(window_len,"d")
else: w = eval("np."+window+"(window_len)")
y = np.convolve(w/w.sum(),s,mode="valid")
d = window_len - 1
y = y[d/2:-d/2]
return y
def get_classifiers():
# basic classifires
dc = DummyClassifier(random_state=0)
lr = LogisticRegression()
gnb = GaussianNB()
svc = LinearSVC(C=1)
C0 = {"name":"Basic","methods":[(dc, "Dummy"), (lr, "Logit"),(gnb, "Naive Bayes"), (svc, "SVC")]}
# decission trees
dec_tree = DecisionTreeClassifier(random_state=0)
etc_tree = ExtraTreeClassifier(random_state=0)
C1 = {"name":"Decision Tree","methods":[(dec_tree, "Decision Tree"),(etc_tree, "Extra Tree")]}
# NN classifirer
knn = KNeighborsClassifier(n_neighbors=25,weights="distance")
rnn = RadiusNeighborsClassifier(radius=20.0,outlier_label=1)
nc = NearestCentroid()
C2 = {"name":"Nearest Neighbors","methods":[(knn, "KNN"),(rnn, "Radius NN"),(nc, "Nearest Centroid")]}
# ensamble models
ada = AdaBoostClassifier()
bg = BaggingClassifier(n_estimators=50, max_features=3)
etsc = ExtraTreesClassifier(n_estimators=50,criterion="entropy")
gb = GradientBoostingClassifier(max_depth=5,random_state=0)
rfc = RandomForestClassifier(n_estimators=100)
C3 = {"name":"Ensemble","methods":[(ada, "Ada Boost"),(bg,"Bagging"),(etsc, "Extra Trees"),
(gb, "Gradient Boosting"), (rfc, "Random Forest")]}
# discriminant analysis & GPC
lda = LinearDiscriminantAnalysis()
qda = QuadraticDiscriminantAnalysis()
C4 = {"name":"Discriminant Analysis","methods":[(lda, "LDA"),(qda, "QDA")]}
# neural net
nn = MLPClassifier(alpha=0.1,tol=1e-8)
C5 = {"name":"Complex Architecture","methods":[(nn, "Neural Network")]}
CLF = [C0,C1,C2,C3,C4,C5]
return CLF
def get_roc_details(clf, X_test, y_test):
if hasattr(clf, "predict_proba"): y_score = clf.predict_proba(X_test)[:, 1]
else:
prob_pos = clf.decision_function(X_test)
y_score = (prob_pos - prob_pos.min()) / (prob_pos.max() - prob_pos.min())
pass
fpr, tpr, threshold = roc_curve(y_test, y_score)
roc_auc = auc(fpr, tpr)
return y_score, fpr, tpr, roc_auc
def plot_deterministic_roc_curves(roc_eval_details, tag):
fig, axes = plt.subplots(nrows=2,ncols=3,figsize=(12,8),dpi=180)
fig.subplots_adjust(wspace=0.5,hspace=0.5)
splot.style("spacepy")
lw = 2
I = 0
for gname in roc_eval_details.keys():
i,j = int(I/3), int(np.mod(I,3))
ax = axes[i,j]
clf_type = roc_eval_details[gname]
for name in clf_type.keys():
roc = roc_eval_details[gname][name]
ax.plot(roc["fpr"], roc["tpr"], color=roc["c"], lw = lw, label="%s:AUC = %0.2f" % (name,roc["roc_auc"]))
pass
ax.plot([0, 1], [0, 1], color="navy", lw=lw, linestyle="--")
ax.set_xlim([0.0, 1.0])
ax.set_ylim([0.0, 1.0])
ax.set_title(gname)
ax.set_xlabel("False Positive Rate")
ax.set_ylabel("True Positive Rate")
ax.legend(loc="lower right",prop={"size": 6})
I = I + 1
pass
fig.savefig("out/deterministinc_forecast_models_roc_curves_%s.png"%tag,bbox_inches="tight")
return
def validate_model_matrices(clf, X_test, y_true):
y_pred = clf.predict(X_test)
CM = confusion_matrix(y_true, y_pred)
C2x2 = Contingency2x2(CM.T)
return C2x2
def get_regressor(name, trw=27):
REGs = {}
# basic regressor
REGs["dummy"] = (DummyRegressor(strategy="median"), name, trw)
REGs["regression"] = (LinearRegression(), name, trw)
REGs["elasticnet"] = (ElasticNet(alpha=.5,tol=1e-2), name, trw)
REGs["bayesianridge"] = (BayesianRidge(n_iter=300, tol=1e-5, alpha_1=1e-06, alpha_2=1e-06, lambda_1=1e-06, lambda_2=1e-06, fit_intercept=True), name, trw)
# decission trees
REGs["dtree"] = (DecisionTreeRegressor(random_state=0,max_depth=5), name, trw)
REGs["etree"] = (ExtraTreeRegressor(random_state=0,max_depth=5), name, trw)
# NN regressor
REGs["knn"] = (KNeighborsRegressor(n_neighbors=25,weights="distance"), name, trw)
# ensamble models
REGs["ada"] = (AdaBoostRegressor(), name, trw)
REGs["bagging"] = (BaggingRegressor(n_estimators=50, max_features=3), name, trw)
REGs["etrees"] = (ExtraTreesRegressor(n_estimators=50), name, trw)
REGs["gboost"] = (GradientBoostingRegressor(max_depth=5,random_state=0), name, trw)
REGs["randomforest"] = (RandomForestRegressor(n_estimators=100), name, trw)
return REGs[name]
def get_hyp_param(kernel_type):
hyp = {}
if kernel_type == "RBF": hyp["l"] = 1.0
if kernel_type == "RQ":
hyp["l"] = 1.0
hyp["a"] = 0.1
if kernel_type == "Matern": hyp["l"] = 1.0
return hyp
def get_gpr(kernel_type, hyp, nrst = 10, trw=27):
if kernel_type == "RBF": kernel = RBF(length_scale=hyp["l"],length_scale_bounds=(1e-02, 1e2))
if kernel_type == "RQ": kernel = RQ(length_scale=hyp["l"],alpha=hyp["a"],length_scale_bounds=(1e-02, 1e2),alpha_bounds=(1e-2, 1e2))
if kernel_type == "Matern": kernel = Matern(length_scale=hyp["l"],length_scale_bounds=(1e-02, 1e2), nu=1.4)
gpr = GaussianProcessRegressor(kernel = kernel, n_restarts_optimizer = nrst)
return (gpr, "GPR", trw)
def get_lstm(ishape,look_back=1, trw = 27):
model = Sequential()
model.add(LSTM(10, input_shape=(look_back, ishape)))
model.add(Dense(1))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='rmsprop')
return (model, "LSTM", trw)
def get_lstm_classifier(ishape):
model = Sequential()
model.add(Embedding(input_dim = 188, output_dim = 50, input_length = ishape))
model.add(LSTM(output_dim=256, activation='sigmoid', inner_activation='hard_sigmoid', return_sequences=True))
model.add(Dropout(0.5))
model.add(LSTM(output_dim=256, activation='sigmoid', inner_activation='hard_sigmoid'))
model.add(Dropout(0.5))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer='rmsprop',metrics=['accuracy'])
return model
def run_validation(pred,obs,year,model):
pred,obs = np.array(pred),np.array(obs)
_eval_details = {}
_eval_details["range"] = "N"
if max(pred) > 9. or min(pred) < 0.: _eval_details["range"] = "Y"
try: _eval_details["bias"] = np.round(verify.bias(pred,obs),2)
except: _eval_details["bias"] = np.NaN
try: _eval_details["meanPercentageError"] = np.round(verify.meanPercentageError(pred,obs),2)
except: _eval_details["meanPercentageError"] = np.NaN
try: _eval_details["medianLogAccuracy"] = np.round(verify.medianLogAccuracy(pred,obs),3)
except: _eval_details["medianLogAccuracy"] = np.NaN
try:_eval_details["symmetricSignedBias"] = np.round(verify.symmetricSignedBias(pred,obs),3)
except: _eval_details["symmetricSignedBias"] = np.NaN
try: _eval_details["meanSquaredError"] = np.round(verify.meanSquaredError(pred,obs),2)
except: _eval_details["meanSquaredError"] = np.NaN
try: _eval_details["RMSE"] = np.round(verify.RMSE(pred,obs),2)
except: _eval_details["RMSE"] = np.NaN
try: _eval_details["meanAbsError"] = np.round(verify.meanAbsError(pred,obs),2)
except: _eval_details["meanAbsError"] = np.NaN
try: _eval_details["medAbsError"] = np.round(verify.medAbsError(pred,obs),2)
except: _eval_details["medAbsError"] = np.NaN
try: _eval_details["nRMSE"] = np.round(verify.nRMSE(pred,obs),2)
except: _eval_details["nRMSE"] = np.NaN
try: _eval_details["forecastError"] = np.round(np.mean(verify.forecastError(pred,obs)),2)
except: _eval_details["forecastError"] = np.NaN
try: _eval_details["logAccuracy"] = np.round(np.mean(verify.logAccuracy(pred,obs)),2)
except: _eval_details["logAccuracy"] = np.NaN
try: _eval_details["medSymAccuracy"] = np.round(verify.medSymAccuracy(pred,obs),2)
except: _eval_details["medSymAccuracy"] = np.NaN
try: _eval_details["meanAPE"] = np.round(verify.meanAPE(pred,obs),2)
except: _eval_details["meanAPE"] = np.NaN
try: _eval_details["medAbsDev"] = np.round(verify.medAbsDev(pred),2)
except: _eval_details["medAbsDev"] = np.NaN
try: _eval_details["rSD"] = np.round(verify.rSD(pred),2)
except: _eval_details["rSD"] = np.NaN
try: _eval_details["rCV"] = np.round(verify.rCV(pred),2)
except: _eval_details["rCV"] = np.NaN
_eval_details["year"] = year
_eval_details["model"] = model
r,_ = pearsonr(pred,obs)
_eval_details["r"] = r
return _eval_details
def get_best_determinsistic_classifier(f_clf):
if not os.path.exists(f_clf):
# Dataset
_xparams,X,y = db.load_data_for_deterministic_bin_clf()
rus = RandomUnderSampler(return_indices=True)
X_resampled, y_resampled, idx_resampled = rus.fit_sample(X, y)
clf = RandomForestClassifier(n_estimators=100)
clf.fit(X_resampled,y_resampled)
joblib.dump(clf, f_clf)
else:
clf = joblib.load(f_clf)
return clf
def get_stats(model, trw):
fname = "out/det.%s.pred.%d.csv"%(model,trw)
fname = "out/det.%s.pred.%d.g.csv"%(model,trw)
print(fname)
_o = pd.read_csv(fname)
_o = _o[(_o.prob_clsf != -1.) & (_o.y_pred != -1.) & (_o.y_pred >= 0) & (_o.y_pred <= 9.)]
y_pred = _o.y_pred.tolist()
y_obs = _o.y_obs.tolist()
_eval_details = run_validation(y_pred,y_obs,"[1995-2016]",model)
print _eval_details
splot.style("spacepy")
fig, ax = plt.subplots(nrows=1,ncols=1,figsize=(6,6))
ax.plot(y_pred,y_obs,"k.")
print("Updated")
strx = "RMSE=%.2f\nr=%.2f"%(_eval_details["RMSE"],_eval_details["r"])
ax.text(0.2,0.95,strx,horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
ax.set_xlabel(r"$K_{P_{pred}}$")
ax.set_xlim(0,9)
ax.set_ylim(0,9)
ax.set_ylabel(r"$K_{P_{obs}}$")
fig.savefig("out/stat/det.%s.pred.%d.png"%(model,trw))
return
def run_for_TSS(model, trw):
fdummy = "out/det.dummy.pred.%d.csv"%(trw)
fname = "out/det.%s.pred.%d.csv"%(model,trw)
_od = pd.read_csv(fdummy)
_o = pd.read_csv(fname)
_od = _od[(_od.prob_clsf != -1.) & (_od.y_pred != -1.) & (_od.y_pred >= 0) & (_od.y_pred <= 9.)]
_o = _o[(_o.prob_clsf != -1.) & (_o.y_pred != -1.) & (_o.y_pred >= 0) & (_o.y_pred <= 9.)]
_od.dn = pd.to_datetime(_od.dn)
_o.dn = pd.to_datetime(_o.dn)
stime = dt.datetime(1995,2,1)
etime = dt.datetime(2016,9,20)
d = stime
skill = []
t = []
while(d < etime):
try:
t.append(d)
dn = d + dt.timedelta(days=27)
dum = _od[(_od.dn >= d) & (_od.dn < dn)]
mod = _o[(_o.dn >= d) & (_o.dn < dn)]
rmse_dum = verify.RMSE(dum.y_pred,dum.y_obs)
rmse = verify.RMSE(mod.y_pred,mod.y_obs)
print(d,rmse,rmse_dum,verify.skill(rmse, rmse_dum))
skill.append(verify.skill(rmse, rmse_dum))
d = d + dt.timedelta(days=1)
except: pass
pass
skill = np.array(skill)
#skill = nan_helper(skill)
fmt = matplotlib.dates.DateFormatter("%d %b\n%Y")
splot.style("spacepy")
fig, ax = plt.subplots(nrows=1,ncols=1,figsize=(10,6))
ax.xaxis.set_major_formatter(fmt)
ax.plot(t,skill,"k.",label="")
#ax.plot(t,smooth(np.array(skill),101),"r.")
#strx = "RMSE:%.2f\nr:%.2f"%(_eval_details["RMSE"],_eval_details["r"])
#ax.text(0.2,0.8,strx,horizontalalignment='center',verticalalignment='center', transform=ax.transAxes)
ax.set_ylabel(r"$TSS(\%)$")
ax.set_xlabel(r"$Time$")
ax.set_xlim(dt.datetime(1995,1,1), dt.datetime(2017,1,1))
ax.set_ylim(0,100)
fig.savefig("out/stat/det.%s.tss.%d.png"%(model,trw))
def plot_pred(model,trw):
fname = "out/det.%s.pred.%d.csv"%(model,trw)
matplotlib.rcParams['xtick.labelsize'] = 10
print(fname)
_o = pd.read_csv(fname)
_o.dn = pd.to_datetime(_o.dn)
_o = _o[(_o.prob_clsf != -1.) & (_o.y_pred != -1.) & (_o.y_pred >= 0) & (_o.y_pred <= 9.)]
_o = _o[(_o.dn >= dt.datetime(2004,7,1)) & (_o.dn <= dt.datetime(2004,8,28))]
_o = _o.drop_duplicates(subset=["dn"])
y_pred = np.array(_o.y_pred.tolist())
y_obs = np.array(_o.y_obs.tolist())
sigma = 3 * np.abs(np.array(_o.y_pred) - np.array(_o.lb))
splot.style("spacepy")
fig, ax = plt.subplots(nrows=1,ncols=1,figsize=(10,6))
fmt = matplotlib.dates.DateFormatter("%m-%d")
ax.xaxis.set_major_formatter(fmt)
ax.plot(_o.dn,y_obs,"ro",markersize=5,label=r"$K_{P_{obs}}$",alpha=0.6)
ax.plot(_o.dn,y_pred,"bo",markersize=3,label=r"$K_{P_{pred}}$")
ax.fill(np.concatenate([_o.dn.tolist(), _o.dn.tolist()[::-1]]),
np.concatenate([y_pred - 1.9600 * sigma,
(y_pred + 1.9600 * sigma)[::-1]]),
alpha=.4, fc='b', ec='None', label='95% confidence interval')
ax.fill(np.concatenate([_o.dn.tolist(), _o.dn.tolist()[::-1]]),
np.concatenate([y_pred - 0.684 * sigma,
(y_pred + 0.684 * sigma)[::-1]]),
alpha=.7, fc='b', ec='None', label='50% confidence interval')
ax.set_ylabel(r"$K_{P_{pred}}$")
ax.set_xlabel(r"$UT$")
ax.legend(loc="upper left")
ax.tick_params(axis="both",which="major",labelsize="15")
ax.set_xlim(dt.datetime(2004,7,1), dt.datetime(2004,8,28))
fig.savefig("out/stat/det.pred.%s.%d.line.png"%(model,trw),bbox_inches="tight")
return
plot_pred("deepGP",27)
def proba_storm_forcast(model,trw):
import matplotlib.gridspec as gridspec
spec = gridspec.GridSpec(ncols=1, nrows=10)
fname = "out/det.%s.pred.%d.csv"%(model,trw)
matplotlib.rcParams['xtick.labelsize'] = 10
print(fname)
_o = pd.read_csv(fname)
_o.dn = pd.to_datetime(_o.dn)
_o = _o[(_o.prob_clsf != -1.) & (_o.y_pred != -1.) & (_o.y_pred >= 0) & (_o.y_pred <= 9.)]
_o = _o[(_o.dn >= dt.datetime(2004,7,22)) & (_o.dn <= dt.datetime(2004,7,28))]
_o = _o.drop_duplicates(subset=["dn"])
y_pred = np.array(_o.y_pred.tolist())
y_obs = np.array(_o.y_obs.tolist())
sigma = 3 * np.abs(np.array(_o.y_pred) - np.array(_o.lb))
splot.style("spacepy")
fig = plt.figure(figsize=(10,6))
fig.subplots_adjust(hspace=0.5)
#fig, ax = plt.subplots(nrows=1,ncols=1,figsize=(10,6))
ax0 = fig.add_subplot(spec[0:2, 0])
ax0.set_ylim(0,9)
ax0.set_yticks([0,3,6,9])
ax0.set_xticks([])
ax0.set_xticklabels([])
ax0.set_xlim(dt.datetime(2004,7,21,21), dt.datetime(2004,7,28,3))
ax = fig.add_subplot(spec[2:, 0])
fmt = matplotlib.dates.DateFormatter("%m-%d")
ax.xaxis.set_major_formatter(fmt)
ax.plot(_o.dn,y_obs,"ro",markersize=5,label=r"$K_{P_{obs}}$",alpha=0.6)
ax.plot(_o.dn,y_pred,"bo",markersize=3,label=r"$K_{P_{pred}}$")
ax.fill(np.concatenate([_o.dn.tolist(), _o.dn.tolist()[::-1]]),
np.concatenate([y_pred - 1.9600 * sigma,
(y_pred + 1.9600 * sigma)[::-1]]),
alpha=.4, fc='b', ec='None', label='95% confidence interval')
ax.fill(np.concatenate([_o.dn.tolist(), _o.dn.tolist()[::-1]]),
np.concatenate([y_pred - 0.684 * sigma,
(y_pred + 0.684 * sigma)[::-1]]),
alpha=.7, fc='b', ec='None', label='50% confidence interval')
ax.plot(_o.dn,4.5*np.ones(len(_o)),"k-.",markersize=3,label=r"$K_{P_{G_0}}$")
ax.set_ylabel(r"$K_{P_{pred}}$")
ax.set_xlabel(r"$UT$")
#ax.legend(loc="upper left")
ax.tick_params(axis="both",which="major",labelsize="15")
ax.set_xlim(dt.datetime(2004,7,21,21), dt.datetime(2004,7,28,3))
cmap = matplotlib.cm.get_cmap('Spectral')
for m,s,d in zip(y_pred, sigma,_o.dn.tolist()):
pr = np.round((1 - norm.cdf(4.5, m, s))*100,1)
c = "g"
if pr > 30.: c = "orange"
if pr > 60.: c = "red"
if pr > 30.: ax.text(d,12.5,str(pr)+"%",rotation=90)
markerline, stemlines, baseline = ax0.stem([d], [m], c)
ax0.plot([d],[m],c,marker="o",markersize=6)
#plt.setp(stemlines, 'color', cmap(pr/100.))
plt.setp(stemlines, 'linewidth', 3.5)
pass
ax.set_ylim(-2,15)
fig.savefig("out/stat/det.pred.%s.%d.forecast.png"%(model,trw),bbox_inches="tight")
return
proba_storm_forcast("deepGP",27)