-
Notifications
You must be signed in to change notification settings - Fork 0
/
ftp_tutorial.py
65 lines (59 loc) · 2.09 KB
/
ftp_tutorial.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
import sys
import pandas as pd
import numpy as np
from sklearn.cross_validation import train_test_split
from sklearn.preprocessing import Imputer
from sklearn.linear_model import Ridge
from sklearn.metrics import mean_squared_error
TEMPERATURE_TRAIN_FEATURE_PATH = 'Temperature_Train_Feature.tsv'
TEMPERATURE_TRAIN_TARGET_PATH = 'Temperature_Train_Target.dat.tsv'
TEMPERATURE_TEST_FEATURE_PATH = 'Temperature_Test_Feature.tsv'
LOCATION = 'Location.tsv'
location = pd.read_csv(LOCATION, sep='\t')
height = location.loc[:, 'height'].values
imp = Imputer(strategy='mean', axis=0)
def train_model(xTrain, yTrain):
# process data
goal = xTrain.shape
h_feat = np.asarray(map(lambda x:height[x] * 0.01, xTrain[:,-1]))
xTrain = np.append(xTrain[:,:-1], h_feat)
xTrain = xTrain.reshape(goal)
sys.exit()
reg = Ridge(alpha=0.1)
reg.fit(xTrain, yTrain)
return reg
### test by splitted data. ###
def model_test():
# read training data.
data_train_feature = pd.read_csv(TEMPERATURE_TRAIN_FEATURE_PATH, sep='\t')
which = ['place%d' % i for i in range(11)] + ['targetplaceid']
X = data_train_feature.loc[:, which].values
y = np.loadtxt(TEMPERATURE_TRAIN_TARGET_PATH)
# split.
TEST_SIZE = 0.2
RANDOM_STATE = 0
X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=TEST_SIZE, random_state=RANDOM_STATE)
# fix nan.
imp.fit(X_train)
X_train = imp.transform(X_train)
imp.fit(X_val)
X_val = imp.transform(X_val)
# train_model
MD = train_model(X_train, y_train)
# predict.
y_val_pred = MD.predict(X_val)
MSE = mean_squared_error(y_val, y_val_pred)
print MSE
def print_submit():
imp.fit(X)
X = imp.transform(X)
MD = train_model(X, y)
data_test_feature = pd.read_csv(TEMPERATURE_TEST_FEATURE_PATH, sep='\t')
X_test = data_test_feature.loc[:, ['place%d' % i for i in range(11)]].values
X_test = imp.transform(X_test)
y_test_pred = MD.predict(X_test)
# print to file.
SUBMIT_PATH = 'submission.dat'
np.savetxt(SUBMIT_PATH, y_test_pred, fmt='%.10f')
if __name__ == "__main__":
model_test()