def test_run():
x, y = get_iris(True)
print_summary(x)
x, y = get_boston()
print_summary(x)
x, y = get_titanic(True)
print_summary(x)
@author: mvanoudh
"""
#import numpy as np
import pandas as pd
from sklearn2.utils import split_xy, print_summary
from sklearn2.datasets import get_titanic_clean
from sklearn2.covariance import compute_correls
pd.options.display.width = 160
df = get_titanic_clean()['Fare Age Sex Embarked Pclass Survived'.split()]
print_summary(df)
def test1():
return compute_correls(df)
def test2():
return compute_correls(df, model='entropy')
def test3():
return compute_correls(df, model='linear')
def test4():
#from sklearn.decomposition import TruncatedSVD
from sklearn2.utils import TransformerWrap, PassThrought, todf
from sklearn2.utils import object_cols, numeric_cols
pd.options.display.width = 160
logging.basicConfig(format='%(asctime)s - %(name)s - %(message)s')
logging.getLogger().setLevel(level=logging.INFO)
seed(0)
x, y = get_titanic(True)
#x, y = get_iris()
print_summary(x)
ppl = Pipeline([
('in', ConstantInputer()), ("da", DateEncoder()),
('en',
FeatureUnion([('nu',
Pipeline([('ft', FunctionTransformer()),
("sc", TransformerWrap(StandardScaler()))])),
('ca',
make_pipeline(FunctionTransformer(), SparseCatEncoder(),
FunctionTransformer()))])),
('fi', make_union(SelectKBest2(), TruncatedSVD2()))
])
params = {
'en__nu__ft__func': lambda x: x[numeric_cols(x)],
def test_date_encoder():
df = get_titanic()
df.fillna(0, inplace=True)
df2 = DateEncoder(ascategory=True).fit_transform(df)
print_summary(df2)
def test_run():
df = pd.DataFrame({'a': [1, 2], 'b': ['bonjour x', 'hello y']})
df['c'] = df.b.apply(lambda s: s.split(' '))
print_summary(df)