Notes

X = pd.read_csv('~/Desktop/Project/wingst3-nov21d1-machinelearning/Bank_Data_Train.csv')
Y = pd.read_csv('~/Desktop/Project/wingst3-nov21d1-machinelearning/Bank_Data_Test.csv')
X['job'] = X['job'].fillna('unknown')

X['education'] =X['education'].fillna('unknown')

X['default'].value_counts()
default_mode = X['default'].mode()[0]
default_mode
X['default'] = X['default'].fillna(default_mode)

balance_avg = round(X['balance'].mean(),0)
balance_avg

X['balance'] = X['balance'].fillna(balance_avg)
X.dropna(subset=['duration'],inplace=True)
X.isnull().sum()
age_m = round(X[X['marital']=='married']['age'].mean(),2)
age_d = round(X[X['marital']=='divorced']['age'].mean(),2)
age_s = round(X[X['marital']=='single']['age'].mean(),2)

def impute(cols):
age = cols[0]
marital = cols[1]
if pd.isnull(age):
if marital == 'married':
return age_m
elif marital == 'divorced':
return age_d
elif marital == 'single':
return age_s
else:
return age
X['age'] = X[['age','marital']].apply(impute,axis=1)
X.select_dtypes(include=['object'])

dict_job={}
count=0
for i in X.job.unique():
dict_job[i] = count
count+=1

dict_marital={}
count=0
for i in X.marital.unique():
dict_marital[i] = count
count+=1

dict_education={}
count=0
for i in X.education.unique():
dict_education[i] = count
count+=1

dict_default={}
count=0
for i in X.default.unique():
dict_default[i] = count
count+=1

dict_housing={}
count=0
for i in X.housing.unique():
dict_housing[i] = count
count+=1
dict_loan={}
count=0
for i in X.loan.unique():
dict_loan[i] = count
count+=1

dict_contact={}
count=0
for i in X.contact.unique():
dict_contact[i] = count
count+=1
dict_month={}
count=0
for i in X.month.unique():
dict_month[i] = count
count+=1

dict_poutcome={}
count=0
for i in X.poutcome.unique():
dict_poutcome[i] = count
count+=1

dict_y={}
count=0
for i in X.y.unique():
dict_y[i] = count
count+=1

X['job'] = X['job'].map(dict_job)
X['marital'] = X['marital'].map(dict_marital)
X['education'] = X['education'].map(dict_education)
X['default'] = X['default'].map(dict_default)
X['housing'] = X['housing'].map(dict_housing)
X['loan'] = X['loan'].map(dict_loan)
X['month'] = X['month'].map(dict_month)
X['contact'] = X['contact'].map(dict_contact)
X['poutcome'] = X['poutcome'].map(dict_poutcome)
X['y'] = X['y'].map(dict_y)
X.select_dtypes(include=['object'])

dict_job={}
count=0
for i in Y.job.unique():
dict_job[i] = count
count+=1

dict_marital={}
count=0
for i in Y.marital.unique():
dict_marital[i] = count
count+=1

dict_education={}
count=0
for i in Y.education.unique():
dict_education[i] = count
count+=1

dict_default={}
count=0
for i in Y.default.unique():
dict_default[i] = count
count+=1

dict_housing={}
count=0
for i in Y.housing.unique():
dict_housing[i] = count
count+=1
dict_loan={}
count=0
for i in Y.loan.unique():
dict_loan[i] = count
count+=1

dict_contact={}
count=0
for i in Y.contact.unique():
dict_contact[i] = count
count+=1
dict_month={}
count=0
for i in Y.month.unique():
dict_month[i] = count
count+=1

dict_poutcome={}
count=0
for i in Y.poutcome.unique():
dict_poutcome[i] = count
count+=1



Y['job'] = Y['job'].map(dict_job)
Y['marital'] = Y['marital'].map(dict_marital)
Y['education'] = Y['education'].map(dict_education)
Y['default'] = Y['default'].map(dict_default)
Y['housing'] = Y['housing'].map(dict_housing)
Y['loan'] = Y['loan'].map(dict_loan)
Y['contact'] = Y['contact'].map(dict_contact)
Y['month'] = Y['month'].map(dict_month)
Y['poutcome'] = Y['poutcome'].map(dict_poutcome)

from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
X = scaler.fit_transform(X)
Y = scaler.fit_transform(Y)
from sklearn.model_selection import train_test_split
#y = X['y']
#X = X.drop(columns=['y'])
X_train, X_test, Y_train, Y_test = train_test_split(X,y,test_size=0.2)
from sklearn.linear_model import LogisticRegression
search_log_reg = LogisticRegression()
search_log_reg.fit(X_train,Y_train)

y_pred1 = search_log_reg.predict(X_test)
from sklearn.metrics import accuracy_score,f1_score
accuracy_score(y_pred1,Y_test)
#f1_score(y_pred1,Y_test)

Logistic_Output = search_log_reg.predict(Y)
Decision_Output
test_data = pd.DataFrame()
test_data['y'] =Logistic_Output
test_data.to_csv('Logistic_Predictions.csv',index=False)