def initialize_model():
'''
Calls on the StatusModel class, processes training data and trains model.
Returns:
StatusModel trained with 2012-2014 3-year loan data. statusmodel class.
'''
model = StatusModel(model=RandomForestRegressor,
parameters={
'n_estimators': 100,
'max_depth': 10
})
try:
df_3c = pd.read_csv('data/LoanStats3c_securev1.csv',
header=True).iloc[:-2, :]
df_3b = pd.read_csv('data/LoanStats3b_securev1.csv',
header=True).iloc[:-2, :]
except (OSError, IOError):
print "Training data not found. Please install from https://www.lendingclub.com/info/download-data.action"
df_train = pd.concat((df_3c, df_3b), axis=0)
df_train = process_features(df_train,
restrict_date=True,
current_loans=True)
model.train_model(df_train)
dump_to_pickle(model, 'pickle/model.pkl')
return model
def initialize_model():
'''
Calls on the StatusModel class, processes training data and trains model.
Returns:
StatusModel trained with 2012-2014 3-year loan data. statusmodel class.
'''
model = StatusModel(model=RandomForestRegressor,
parameters={'n_estimators':100,
'max_depth':10})
try:
df_3c = pd.read_csv('data/LoanStats3c_securev1.csv', header=True).iloc[:-2, :]
df_3b = pd.read_csv('data/LoanStats3b_securev1.csv', header=True).iloc[:-2, :]
except (OSError, IOError):
print "Training data not found. Please install from https://www.lendingclub.com/info/download-data.action"
df_train = pd.concat((df_3c, df_3b), axis=0)
df_train = process_features(df_train, restrict_date=True, current_loans=True)
model.train_model(df_train)
dump_to_pickle(model, 'pickle/model.pkl')
return model
def test_expected_current():
print "Loading data..."
try:
df_3c = pd.read_csv('data/LoanStats3c_securev1.csv',
header=True).iloc[:-2, :]
df_3b = pd.read_csv('data/LoanStats3b_securev1.csv',
header=True).iloc[:-2, :]
except OSError, IOError:
print "Training data not found. Please install from https://www.lendingclub.com/info/download-data.action"
df_raw = pd.concat((df_3c, df_3b), axis=0)
print "Pre-processing data..."
df = process_features(df_raw)
print "Initializing model..."
model = StatusModel(model=RandomForestRegressor,
parameters={
'n_estimators': 100,
'max_depth': 10
})
print "Training model..."
try:
model = load_from_pickle('pickle/model.pkl')
except OSError, IOError:
print "Model not found. Training model, this might take some time..."
model.train_model(df)
dump_to_pickle(model, 'pickle/model.pkl')
from model.start import initialize_model
from model.validate import actual_IRR
def test_expected_current():
print "Loading data..."
try:
df_3c = pd.read_csv("data/LoanStats3c_securev1.csv", header=True).iloc[:-2, :]
df_3b = pd.read_csv("data/LoanStats3b_securev1.csv", header=True).iloc[:-2, :]
except OSError, IOError:
print "Training data not found. Please install from https://www.lendingclub.com/info/download-data.action"
df_raw = pd.concat((df_3c, df_3b), axis=0)
print "Pre-processing data..."
df = process_features(df_raw)
print "Initializing model..."
model = StatusModel(model=RandomForestRegressor, parameters={"n_estimators": 100, "max_depth": 10})
print "Training model..."
try:
model = load_from_pickle("pickle/model.pkl")
except OSError, IOError:
print "Model not found. Training model, this might take some time..."
model.train_model(df)
dump_to_pickle(model, "pickle/model.pkl")
print "Calculating IRR..."
int_rate_dict = {
"A1": 0.0603,
print "Inserting results of API request to database..."
try:
insert_into_mongodb(loan_results, loan_details)
except Exception:
print "MongoDB error, proceeding to next step..."
print "Loading model..."
try:
model = load_from_pickle('pickle/model.pkl')
except OSError, IOError:
print "Model not found. Initializing training process, this might take some time..."
model = initialize_model()
print "Pre-processing data..."
df_raw = process_requests(loan_results, loan_details)
df = process_features(df_raw, restrict_date=False, features_dict=model.features_dict)
print "Calculating results for display..."
df_display = process_for_display(model, df, loan_results)
print "Inserting processed data to database..."
try:
insert_into_postgresql(df_display)
except Exception:
print "PostgreSQL error, proceeding to next step..."
print "Generating data for charts..."
df_max = df_display.groupby('sub_grade').max()['IRR']
generate_for_charts(df_max)
print "Reformatting for display..."
