Newer
Older
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
{
"cells": [
{
"cell_type": "markdown",
"id": "a5b7fa31",
"metadata": {},
"source": [
"# DRAVOGRAD- DRAVA RIVER\n",
"\n",
"19/10/2021\n",
"\n",
"In this notebook results of different feature selections are compared for the Dravograd Basin (of which we have around 40 years of data)\n",
"\n",
"Input data are ERA5 metereological reanalysis quantile mapped and downscaled by ZAMG.\n",
"\n",
"15 days averages over the previous year of pecipitation, temperature and potential evapotranspiration are selected as input.\n",
"\n",
"The settings are the following:\n",
"\n",
" A) 180 features are selected with PCA, the same numeriosity as setting C) ;\n",
"\n",
" B) 36 features are selectedwith PCA, the same numeriosity as setting D) ;\n",
" \n",
" C) metereological inputs spatial statistics are used as input: mean, the 5th, 25th, 75th and 95th quantiles are selected.\n",
" \n",
" D) metereological inputs are spatially averaged.\n"
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "efcc49ff",
"metadata": {},
"outputs": [],
"source": [
"from sf_runoff import daily_climatology, spatial_avg_daily_input, spatial_stats_daily_input, compute_anomalies\n",
"from nested_CV import SVR_nested_CV_gridsearch, SVR_PCA_nested_CV_gridsearch\n",
"from test import evaluate_prediction, plot_prediction, plot_anomalies\n",
"from test import nested_CV_PCA_SVR_predict, nested_CV_SVR_predict\n",
"from classic_CV_predict import classic_CV_PCA_SVR_predict, classic_CV_SVR_predict\n",
"\n",
"\n",
"import pandas as pd\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import pandas as pd\n",
"import numpy as np\n",
"from scipy.stats import gaussian_kde\n",
"\n",
"from sklearn.svm import SVR\n",
"from sklearn.preprocessing import StandardScaler\n",
"from sklearn.pipeline import make_pipeline\n",
"from sklearn.compose import TransformedTargetRegressor\n",
"from sklearn.model_selection import GridSearchCV,TimeSeriesSplit\n",
"from sklearn.metrics import mean_squared_error\n",
"from sklearn.decomposition import PCA\n",
"\n",
"\n",
"import matplotlib.pyplot as plt\n",
"\n",
"import os\n",
"\n",
"import pdb\n",
"import seaborn as sns"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "f11761aa",
"metadata": {},
"outputs": [],
"source": [
"path=r'C:\\Users\\mmazzolini\\OneDrive - Scientific Network South Tyrol\\Documents\\conda\\daily_input\\\\'\n",
"\n",
"daily_input = pd.read_csv(path+'HEDravograd_Drava_1952_2019.txt', index_col=0, parse_dates=True).astype('float32')\n",
"\n",
"daily_input_TPE = spatial_avg_daily_input(daily_input).astype('float32')\n",
"\n",
"daily_input_stat = spatial_stats_daily_input(daily_input).astype('float32')"
]
},
{
"cell_type": "markdown",
"id": "bab54ce7",
"metadata": {},
"source": [
"import sys, importlib\n",
"importlib.reload(sys.modules['test'])\n"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "a024e5fe",
"metadata": {},
"outputs": [],
"source": [
"#define the possible parameters value (where Gridsearch is applied)\n",
"\n",
"C_range=np.logspace(-2, 2, 10)\n",
"epsilon_range=np.logspace(-6, -2, 5)\n",
"#n_range = [17, 50, 200]\n",
"components_range = [5*3*24]\n",
"#do not enlarge t_range for now\n",
"\n",
"t_range=[24]\n",
"\n",
"#define the temporal unit\n",
"t_unit=15\n",
"n_splits=10\n",
"test_size=365"
]
},
{
"cell_type": "markdown",
"id": "e7d5c48a",
"metadata": {},
"source": [
"# A) PCA+SVR"
]
},
{
"cell_type": "markdown",
"id": "18861993",
"metadata": {},
"source": [
"### TRAIN A PCA+SVR MODEL "
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "aacb3a01",
"metadata": {
"scrolled": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"C:\\Users\\mmazzolini\\.conda\\envs\\ado\\lib\\site-packages\\joblib\\externals\\loky\\process_executor.py:688: UserWarning: A worker stopped while some jobs were given to the executor. This can be caused by a too short worker timeout or by a memory leak.\n",
" warnings.warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n",
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"C:\\Users\\mmazzolini\\.conda\\envs\\ado\\lib\\site-packages\\sklearn\\model_selection\\_search.py:922: UserWarning: One or more of the test scores are non-finite: [0.56732319 0.56730931 0.56738403 0.56741894 0.56698353 0.64647372\n",
" 0.64641015 0.64642175 0.64644063 0.64708145 0.6675348 0.66753546\n",
" 0.66758763 0.6680089 0.66868631 0.66593785 0.66580957 0.6661129\n",
" 0.66658101 0.66807678 0.6631981 0.66284985 0.66283214 0.66275594\n",
" 0.66434368 0.65481724 0.65625838 0.65489764 0.65542396 0.65536472\n",
" 0.63989756 0.64083217 0.64333993 0.64097485 nan 0.64402253\n",
" nan 0.64316301 0.64565183 0.64709809 0.65608431 0.66025241\n",
" nan nan nan nan nan 0.64122172\n",
" nan 0.62904516]\n",
" warnings.warn(\n",
"C:\\Users\\mmazzolini\\.conda\\envs\\ado\\lib\\site-packages\\sklearn\\model_selection\\_search.py:922: UserWarning: One or more of the train scores are non-finite: [0.45772834 0.45772035 0.45772763 0.45770001 0.45719031 0.58781645\n",
" 0.5878289 0.58779987 0.58776526 0.58765745 0.70777745 0.70774931\n",
" 0.70775225 0.70778127 0.70794423 0.81819123 0.81816976 0.81811986\n",
" 0.81820455 0.81818229 0.91899307 0.91895802 0.91898942 0.91900283\n",
" 0.9190149 0.9790693 0.97906333 0.97907522 0.97905555 0.97908029\n",
" nan 0.99416289 0.99415837 0.99418102 nan 0.9981007\n",
" nan 0.9981009 0.99810992 0.9981407 0.99948531 0.99948513\n",
" nan nan nan nan nan 0.99992782\n",
" nan 0.99985103]\n",
" warnings.warn(\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Fitting 1 folds for each of 50 candidates, totalling 50 fits\n"
]
},
Loading
Loading full blame...