-
Notifications
You must be signed in to change notification settings - Fork 3
/
manuscript_results.py
276 lines (216 loc) · 9.97 KB
/
manuscript_results.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
import json
import random
import pickle
import time
import sys
import math
import copy
import os.path
import pandas as pd
import numpy as np
import argparse
import itertools
from data.data import Data
from node.node import Node
def tree_pprinter(node):
fmtr = ""
def pprinter(anode):
nonlocal fmtr
if anode.split_var is not None:
print("({0}) {1} {2}".format(len(anode.data.df.index), anode.split_var, anode.data.var_desc[anode.split_var]['bounds']))
else:
print("({0}) Leaf {1:.2f} {2}".format(len(anode.data.df.index), np.var(anode.data.df[anode.data.class_var].values),
["{} {}".format(key, anode.data.var_desc[key]['bounds']) for key in anode.data.var_desc.keys() if anode.data.var_desc[key]['bounds'] != [[-np.inf, np.inf]]]))
if anode.left_child is not None:
print("{} `--".format(fmtr), end="")
fmtr += " | "
pprinter(anode.left_child)
fmtr = fmtr[:-4]
print("{} `--".format(fmtr), end="")
fmtr += " "
pprinter(anode.right_child)
fmtr = fmtr[:-4]
return pprinter(node)
def tree_trainer(df, class_var, var_desc, stop=50, variance=.001):
if class_var not in df.columns:
raise Exception('Class variable not in DataFrame')
data = Data(df, class_var, var_desc)
node = Node(data, stop=stop, variance=variance)
node.split()
return node
def is_in_bounds(bounds, value):
for bound in bounds:
if bound[0] > bound[1]:
if bound[0] <= value <= 360.0 or 0.0 <= value < bound[1]:
return True
elif bound[0] == 0.0 and value == 360:
return True
else:
if bound[0] == 0.0 and value == 360:
return True
elif bound[0] <= value < bound[1]:
return True
return False
def tree_eval(node, row):
result = None
def eval(node, row):
nonlocal result
if node.split_var is None:
result = np.mean(node.data.df[node.data.class_var].values)
else:
if is_in_bounds(node.left_child.data.var_desc[node.split_var]["bounds"], row[node.split_var]):
eval(node.left_child, row)
elif is_in_bounds(node.right_child.data.var_desc[node.split_var]["bounds"], row[node.split_var]):
eval(node.right_child, row)
else:
print(node.data.var_desc[node.split_var]["bounds"], row[node.split_var])
print(node.left_child.data.var_desc[node.split_var]["bounds"], row[node.split_var])
print(node.right_child.data.var_desc[node.split_var]["bounds"], row[node.split_var])
eval(node, row)
return result
def cxval_k_folds_split(df, k_folds, seed=1):
random.seed(seed)
dataframes = []
group_size = int(round(df.shape[0]*(1.0/k_folds)))
for i in range(k_folds-1):
rows = random.sample(list(df.index), group_size)
dataframes.append(df.ix[rows])
df = df.drop(rows)
dataframes.append(df)
return dataframes
def cxval_select_fold(i_fold, df_folds):
df_folds_copy = copy.deepcopy(df_folds)
if 0 <= i_fold < len(df_folds):
test_df = df_folds_copy[i_fold]
del df_folds_copy[i_fold]
train_df = pd.concat(df_folds_copy)
return train_df, test_df
else:
raise Exception('Group not in range!')
def cxval_test(df_folds, class_var, var_desc, leaf_size):
rmse_results = []
ia_results = []
ia2_results = []
for i in range(len(df_folds)):
train_df, test_df = cxval_select_fold(i, df_folds)
tree = tree_trainer(train_df, class_var, var_desc, leaf_size)
rmse_results.append(tree_rmse_calc(tree, test_df))
ia_results.append(tree_ia_calc(tree, test_df))
ia2_results.append(tree_ria_calc(tree, test_df))
return sum(rmse_results)/len(df_folds), sum(ia_results)/len(df_folds), sum(ia2_results)/len(df_folds)
def tree_rmse_calc(node, df):
acc = 0.0
total_len = len(df.index)
for _, row in df.iterrows():
acc += math.pow((tree_eval(node, row) - row[node.data.class_var]), 2)
return math.sqrt(acc / total_len)
def tree_ia_calc(node, df):
total_len = len(df.index)
sim = np.zeros(total_len)
df.reset_index(drop=True, inplace=True)
for i, row in df.iterrows():
sim[i] = tree_eval(node, row)
obs = df[node.data.class_var]
return 1 - np.sum(np.square(obs - sim)) / np.sum(np.square(np.abs(sim - np.mean(obs)) + np.abs(obs - np.mean(obs))))
def tree_ria_calc(node, df):
total_len = len(df.index)
sim = np.zeros(total_len)
df.reset_index(drop=True, inplace=True)
for i, row in df.iterrows():
sim[i] = tree_eval(node, row)
obs = df[node.data.class_var]
first = np.sum(np.abs(obs - sim))
second = (2*np.sum(np.abs(obs - np.mean(obs))))
if first <= second:
return 1 - first/second
else:
return second/first - 1
def config_generator():
outputs = ["metar_wind_spd", "metar_temp", "metar_rh"]
inputs_lin = ["gfs_wind_spd", "gfs_temp", "gfs_rh"]
inputs_cir = ["gfs_wind_dir", "time", "date"]
for output in outputs:
for l in range(1, len(inputs_lin)+1):
for subset_lin in itertools.combinations(inputs_lin, l):
if not "gfs"+output[5:] in subset_lin:
continue
for subset_cir in itertools.combinations(inputs_cir, l):
config = {}
config["output"] = output
config["input"] = []
for var in subset_lin:
var_desc = {}
var_desc["name"] = var
var_desc["type"] = "lin"
config["input"].append(var_desc)
for var in subset_cir:
var_desc = {}
var_desc["name"] = var
var_desc["type"] = "cir"
config["input"].append(var_desc)
yield config
if __name__ == "__main__":
if len(sys.argv) == 1:
print("Please specify a tree configuration as an argument to this script")
sys.exit()
with open(sys.argv[1]) as conf_file:
tree_params = json.load(conf_file)
class_var = tree_params['output']['name']
tree_desc_lin = {}
for var in tree_params['input']:
tree_desc_lin[var['name']] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_lund = {}
for var in tree_params['input']:
if var["type"] == "cir":
tree_desc_lund[var['name']] = {"type": var['type'], "method": "non-cont", "bounds": [[-np.inf, np.inf]]}
else:
tree_desc_lund[var['name']] = {"type": var['type'], "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_cir = {}
for var in tree_params['input']:
tree_desc_cir[var['name']] = {"type": var['type'], "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_uv = {}
for var in tree_params['input']:
if var['name'] == "gfs_wind_dir":
tree_desc_uv["u_speed"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_uv["v_speed"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
if var['name'] == "time":
tree_desc_uv["u_time"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_uv["v_time"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
if var['name'] == "date":
tree_desc_uv["u_date"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
tree_desc_uv["v_date"] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
if var['type'] == "lin" and var['name'] != "gfs_wind_spd":
tree_desc_uv[var['name']] = {"type": "lin", "method": "cont", "bounds": [[-np.inf, np.inf]]}
#print(tree_desc_lin)
#print(tree_desc_lund)
#print(tree_desc_cir)
#print(tree_desc_uv)
print("")
print("")
print("| Airport | Method | 1000 | 500 | 250 | 100 | 50 |")
print("| ------- |----------| ------:|-------:| ------:|-------:| ------:|")
data_paths = ["datasets/eddt.csv", "datasets/egll.csv", "datasets/lebl.csv", "datasets/lfpg.csv", "datasets/limc.csv"]
airport_codes = ['EDDT', 'EGLL', 'LEBL', 'LFPG', 'LIMC']
for ds_idx, dataset in enumerate(data_paths):
df = pd.read_csv(dataset)
for var in tree_params['input']:
if var['type'] == "cir":
df[var['name']].replace([360], 0, inplace=True)
df["v_speed"] = df["gfs_wind_spd"] * np.sin(np.deg2rad(df["gfs_wind_dir"]))
df["u_speed"] = df["gfs_wind_spd"] * np.cos(np.deg2rad(df["gfs_wind_dir"]))
df["v_time"] = np.sin(np.deg2rad(df["time"]))
df["u_time"] = np.cos(np.deg2rad(df["time"]))
df["v_date"] = np.sin(np.deg2rad(df["date"]))
df["u_date"] = np.cos(np.deg2rad(df["date"]))
df_folds = cxval_k_folds_split(df, 5, 0)
print("| {} |".format(airport_codes[ds_idx]), end='')
tree_names = ["Linear", "Lund", "Circular", "UV"]
for tree_idx, tree_desc in enumerate([tree_desc_lin, tree_desc_lund, tree_desc_cir, tree_desc_uv]):
if tree_idx == 0:
print(" {0: <8} |".format(tree_names[tree_idx]), end='')
else:
print("| | {0: <8} |".format(tree_names[tree_idx]), end='')
for size in [1000, 500, 250, 100, 50]:
rmse, ia, ria = cxval_test(df_folds, class_var, tree_desc, size)
print(" {0:0.4f} |".format(ria), end='')
print("", flush=True)