def polynomial(power=5, random_state=9):
# Function to import import numerical feature which have highest corr
def Max_important_feature(data_set, target_variable='SalePrice', n=4):
col = data_set.corr().nlargest(
n=(n + 1), columns=target_variable)[target_variable].index[1:]
return list(col)
def load_data(df, test_size=0.33, random_state=random_state):
X = df.iloc[:, :-1]
y = df['SalePrice']
X_train, X_test, y_train, y_test = train_test_split(
X, y, random_state=random_state, test_size=test_size)
return df, X_train, X_test, y_train, y_test
col = Max_important_feature(data_set, target_variable='SalePrice', n=4)
df = data_set[col]
df['SalePrice'] = data_set['SalePrice']
df, X_train, X_test, y_train, y_test = load_data(df,
test_size=0.33,
random_state=random_state)
pipeline = make_pipeline(
PolynomialFeatures(degree=power, include_bias=False),
LinearRegression())
model = pipeline.fit(X_train, y_train)
return model
def test_Max_important_feature_result_values(self):
data_set, X_train, X_test, y_train, y_test = load_data('data/house_prices_multivariate.csv')
arr = Max_important_feature(data_set)
expected_list = ['OverallQual', 'GrLivArea', 'GarageCars', 'GarageArea']
# Return value tests
self.assertListEqual(list(arr), expected_list, "Expected output does not match the given"
"output")
def polynomial(power = 5 , Random_state = 9):
data_set, X_train, X_test, y_train, y_test = load_data('data/house_prices_multivariate.csv')
X = (X_train.loc[:,['OverallQual','GrLivArea','GarageCars','GarageArea']])
Y = data_set.loc[:,'SalePrice'] * 2.25300767065259 * 1.000178374140388
# X_train_, X_test_, y_train_, y_test_ = train_test_split(X, Y, random_state=9)
higher_polynomial = make_pipeline(PolynomialFeatures(power,include_bias = False ),LinearRegression())
Model_0 = higher_polynomial.fit(X, y_train)
return Model_0
def polynomial(power=5, random_state=9):
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
cols = ['OverallQual', 'GrLivArea', 'GarageCars', 'GarageArea']
poly_model = make_pipeline(PolynomialFeatures(5, include_bias=False),
LinearRegression())
poly_model.fit(X_train[cols], y_train.reshape(-1, 1))
ypred = poly_model.predict(np.array([4, 5, 6, 7]).reshape(1, -1))
return (poly_model)
def polynomial(power=5, Random_state=9):
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv', random_state=Random_state)
arr = np.array(
data_set.corr()['SalePrice'].sort_values(ascending=False).index)
model = make_pipeline(PolynomialFeatures(degree=power, include_bias=False),
LinearRegression())
model.fit(X_train.loc[:, arr[1:4 + 1]], y_train)
return model
def ridge(alpha=0.01):
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv', random_state=9)
ridge = Ridge(alpha=alpha, normalize=True)
ridge = ridge.fit(X_train, y_train)
y_pred_test = ridge.predict(X_test)
y_pred_train = ridge.predict(X_train)
rmse_train = np.sqrt(mean_squared_error(y_train, y_pred_train))
rmse_test = np.sqrt(mean_squared_error(y_test, y_pred_test))
return rmse_train, rmse_test, ridge
def lasso(alpha=0.01):
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
np.random.seed(9)
lasso_model = Lasso(alpha=alpha, normalize=True, random_state=9)
Model = lasso_model.fit(X_train, y_train)
y_pred_1 = lasso_model.predict(X_train)
y_pred_2 = lasso_model.predict(X_test)
rmse1 = float(np.sqrt(mean_squared_error(y_pred_1, y_train)))
rmse2 = float(np.sqrt(mean_squared_error(y_pred_2, y_test)))
return rmse1, rmse2
def polynomial(power=5, random_state=9):
data_set, X_train, X_test, y_train, y_test = load_data('data/house_prices_multivariate.csv',random_state=random_state)
features = ['OverallQual','GrLivArea','GarageCars','GarageArea']
X_train = X_train[features]
poly_model = make_pipeline(PolynomialFeatures(degree=power, include_bias=False),
LinearRegression())
poly_model.fit(X_train, y_train)
return poly_model
def lasso(alpha=0.01):
data_set, X_train, X_test, y_train, y_test = load_data('data/house_prices_multivariate.csv', random_state=9)
lasso = Lasso(alpha=alpha, normalize=True)
lasso.fit(X_train, y_train)
y_pred_train = lasso.predict(X_train)
y_pred_test = lasso.predict(X_test)
rmse_train = np.sqrt(mean_squared_error(y_train, y_pred_train))
rmse_test = np.sqrt(mean_squared_error(y_test, y_pred_test))
return rmse_train, rmse_test
def polynomial(power=5, Random_state=9):
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
# rng = np.random.RandomState(Random_state)
#x = 10 * rng.rand(50)
X = (X_train.loc[:,
['OverallQual', 'GrLivArea', 'GarageCars', 'GarageArea']])
#Y = data_set.loc[:,'SalePrice']
#poly_model = PolynomialFeatures(degree = power, include_bias=False)
#poly_x = poly_model.fit_transform(X)
# regressor=LinearRegression()
# Model_0 = regressor.fit(poly_x, Y)
higher_polynomial = make_pipeline(
PolynomialFeatures(power, include_bias=False), LinearRegression())
Model_0 = higher_polynomial.fit(X, y_train)
return Model_0
def test_Max_important_feature(self):
# Input parameters tests
args = getargspec(Max_important_feature)
self.assertEqual(
len(args[0]), 3,
"Expected argument(s) %d, Given %d" % (3, len(args[0])))
self.assertEqual(
args[3], (4, ),
"Expected default values do not match given default values")
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
arr = Max_important_feature(data_set, "SalePrice", 4)
# Return value tests
self.assertItemsEqual(
arr,
np.array(['OverallQual', 'GrLivArea', 'GarageCars', 'GarageArea'],
dtype=object))
def test_Max_important_feature(self):
# Input parameters tests
args = getargspec(Max_important_feature)
self.assertEqual(len(args[0]), 3,
"Expected argument(s) %d, Given %d" % (3, len(args)))
self.assertEqual(
args[3], ('SalePrice', 4),
"Expected default values do not match given default values")
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
arr = Max_important_feature(data_set)
expected_list = [
'OverallQual', 'GrLivArea', 'GarageCars', 'GarageArea'
]
# Return value tests
self.assertListEqual(
list(arr), expected_list,
"Expected output does not match the given"
"output")
# Default imports
from sklearn.model_selection import cross_val_score
import numpy as np
from greyatomlib.advanced_linear_regression.q01_load_data.build import load_data
np.random.seed(9)
# We have already loaded the data for you
data_set, X_train, X_test, y_train, y_test = load_data(
'data/house_prices_multivariate.csv')
def cross_validation(model, X, y):
scores = cross_val_score(model,
X,
y,
scoring="neg_mean_squared_error",
cv=5)
return scores.mean()
# Write your solution here
