def test_raises_error_if_df_contains_na(df_enc_big, df_enc_big_na):
# test case 4: when dataset contains na, fit method
with pytest.raises(ValueError):
encoder = OneHotEncoder()
encoder.fit(df_enc_big_na)
# test case 4: when dataset contains na, transform method
with pytest.raises(ValueError):
encoder = OneHotEncoder()
encoder.fit(df_enc_big)
encoder.transform(df_enc_big_na)
def test_get_feature_names_out(df_enc_binary):
original_features = ["var_num"]
input_features = ["var_A", "var_B", "var_C", "var_D"]
tr = OneHotEncoder()
tr.fit(df_enc_binary)
out = [
"var_A_A",
"var_A_B",
"var_A_C",
"var_B_A",
"var_B_B",
"var_B_C",
"var_C_AHA",
"var_C_UHU",
"var_D_OHO",
"var_D_EHE",
]
assert tr.get_feature_names_out(
input_features=None) == original_features + out
assert tr.get_feature_names_out(input_features=input_features) == out
assert tr.get_feature_names_out(
input_features=input_features[0:2]) == out[0:6]
assert tr.get_feature_names_out(
input_features=[input_features[0]]) == out[0:3]
tr = OneHotEncoder(drop_last=True)
tr.fit(df_enc_binary)
out = [
"var_A_A",
"var_A_B",
"var_B_A",
"var_B_B",
"var_C_AHA",
"var_D_OHO",
]
assert tr.get_feature_names_out(
input_features=None) == original_features + out
assert tr.get_feature_names_out(input_features=input_features) == out
assert tr.get_feature_names_out(
input_features=input_features[0:2]) == out[0:4]
assert tr.get_feature_names_out(
input_features=[input_features[0]]) == out[0:2]
tr = OneHotEncoder(drop_last_binary=True)
tr.fit(df_enc_binary)
out = [
"var_A_A",
"var_A_B",
"var_A_C",
"var_B_A",
"var_B_B",
"var_B_C",
"var_C_AHA",
"var_D_OHO",
]
assert tr.get_feature_names_out(
input_features=None) == original_features + out
assert tr.get_feature_names_out(input_features=input_features) == out
assert tr.get_feature_names_out(
input_features=[input_features[0]]) == out[0:3]
assert tr.get_feature_names_out(input_features=[input_features[3]]) == [
out[-1]
]
tr = OneHotEncoder(top_categories=1)
tr.fit(df_enc_binary)
out = ["var_A_B", "var_B_A", "var_C_AHA", "var_D_EHE"]
assert tr.get_feature_names_out(
input_features=None) == original_features + out
assert tr.get_feature_names_out(input_features=input_features) == out
assert tr.get_feature_names_out(
input_features=input_features[0:2]) == out[0:2]
assert tr.get_feature_names_out(input_features=[input_features[3]]) == [
out[3]
]
with pytest.raises(ValueError):
tr.get_feature_names_out("var_A")
with pytest.raises(ValueError):
tr.get_feature_names_out(["var_A", "hola"])
st.selectbox('Do you want the house to be furnished ?', ('Yes', 'No')))
security_doors = word_convert(
st.selectbox('Do you want security doors ?', ('Yes', 'No')))
cctv = word_convert(
st.selectbox('Do you want CCTV surveillance ?', ('Yes', 'No')))
bq = word_convert(st.selectbox('Do you want Boys Quarters ?', ('Yes', 'No')))
gym = word_convert(st.selectbox('Do you need gym facilities ?', ('Yes', 'No')))
pool = word_convert(st.selectbox('Do you need swimming pool ?', ('Yes', 'No')))
# Modeling step
# Encoding Step
encode = OneHotEncoder()
target = data['Price']
features = data.drop('Price', 1)
encode.fit(features)
features = encode.transform(features)
# Getting the target and features variables
# print(data.head())
X_train, X_test, y_train, y_test = train_test_split(features,
target,
test_size=0.2,
random_state=0)
# Creating the algorithm class
model = RandomForestRegressor()
# Creating algorithm object
model.fit(X_train, y_train)
# Predicted values
# Criando um modelo com todas as features e usando pipeline
num_features = df.select_dtypes(include=['int64', 'float64']).drop(
'Survived', axis=1).columns
num_features
# %%
cat_features = df.select_dtypes(include=['category', 'object']).columns
cat_features
#%%
features = df.drop('Survived', axis=1).columns.to_list()
features
# %%
onehot = OneHotEncoder(variables=['Pclass', 'Sex', 'Embarked'],
drop_last=False)
# %%
onehot.fit(df[features])
onehot.transform(df[features]).head()
# %%
X = onehot.transform(df[features])
y = df['Survived']
print(X.shape)
# %%
# Separate into train and test sets
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.2,
random_state=0)
logistic_model = LogisticRegression(penalty='l2', C=1.0,
solver='liblinear').fit(X_train, y_train)
print(logistic_model)
def encorder(self, y):
"""Y dataframe"""
encode = OneHotEncoder()
encode.fit(y)
return encode.transform(y)