from scipy.stats import chi2
from scipy.stats import entropy
import numpy as np
from collections import Counter
import math
from petfinder.get_explore import read_data, Columns
from petfinder.preprocessing import prepare_data
from petfinder.feature_engineering import finalize_data, add_features, filter_by_varth
from petfinder.tools import detect_outliers, plog
if __name__ == "__main__":
sys.stdout.buffer.write(chr(9986).encode('utf8'))
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1000)
train, test = read_data()
train, test = prepare_data(train, test)
train, test = add_features(train, test)
ccols = train.columns.values.tolist()
for c in ccols:
if (train[c].isna().any()) & \
(c in Columns.img_num_cols_1.value + Columns.img_num_cols_2.value + Columns.img_num_cols_3.value ):
plog("filling " + c + " with 0 on train dataset")
train[c].fillna(0, inplace=True)
ccols = test.columns.values.tolist()
for c in ccols:
if (test[c].isna().any()) & \
(c in Columns.img_num_cols_1.value + Columns.img_num_cols_2.value + Columns.img_num_cols_3.value ):
plog("filling " + c + " with 0 on test dataset")
test[c].fillna(0, inplace=True)
from mord import OrdinalRidge
from petfinder.get_explore import read_data
from petfinder.preprocessing import prepare_data
import pandas as pd
if __name__ == "__main__":
train, test = read_data()
x_train, y_train, x_test, id_test = prepare_data(train, test)
clf = OrdinalRidge()
clf.fit(x_train, y_train)
pred = clf.predict(x_test)
print(clf.score(x_train, y_train))
print(pd.DataFrame({'PetID': id_test.PetID.values.ravel(), 'AdoptionSpeed': pred}))
'PetID': test_id.values.ravel(),
'AdoptionSpeed': pred
})
return prediction_df
if __name__ == "__main__":
pd.set_option('display.max_columns', 500)
pd.set_option('display.width', 1500)
df = pd.DataFrame(np.random.randn(5, 3), columns=['perf', 'B', 'C'])
#predict(df)
train, test = read_data()
train_x, train_y, test_x, test_id = prepare_data(train, test)
#rs = randomsearchpipeline(train_x.drop(["RescuerID"], axis=1), train_y)
train_x.drop(["RescuerID"], axis=1, inplace=True)
test_x.drop(["RescuerID"], axis=1, inplace=True)
#print(train_x)
#print(rs)
#sys.exit()
#print(train_x)
#print(train_y)
#print(sp.stats.randint(1, 6).value)
clf, mean = iterate_by_randomsearch(train_x, train_y.values.ravel())
print(clf)
print(mean)
#for c in clf[""]:
#print(c)