def fit_and_predict(df,
outcome: str = 'deaths',
method: str = 'exponential',
mode: str = 'predict_future',
target_day: np.ndarray = np.array([1]),
output_key: str = None,
demographic_vars=[]):
"""
Trains a method (method) to predict a current number of days ahead (target_day)
Predicts the values of the number of deaths for the final day of test_df and writes to the column
'predicted_deaths_'+method+'_'+str(target_day[-1]) of the test_df
Params
------
df
a df with county level deaths and cases and demographic information
outcome
key for the outcome to predict (the values in this column should have a list for each row)
method
what method to use to do forecasting
target_day
np.array([1,2,..,n]) predicts these number of days ahead (can just be np.array([3])) for example if you just want 3 days ahead)
output_key
key to save the output as
mode:
either 'predict_future' or 'eval_mode'
predict_future is predicting deaths on FUTURE days, so target_day=np.array([1])) means it predicts tomorrow's deaths
eval_mode is for evaluating the performance of the classifier.
target_day=np.array([k])) will predict the current days death count using information from k days ago.
target_day= np.array([1,2,3,...,k]) will predict todays deaths, yesterdays deaths, deaths k-1 days ago using information from k days ago.
Returns
-------
test_df
returns dataframe with added column
"""
assert mode == 'predict_future' or mode == 'eval_mode', 'unknown mode'
if output_key is None:
output_key = f'predicted_{outcome}_{method}_{target_day[-1]}'
if len(demographic_vars) > 0:
output_key += '_demographics'
if method == 'AR':
print('currently deprecated')
raise NotImplementedError
loss, model, best_window = naive_autoreg_baselines.train_and_evaluate_model(
train_df, test_df)
return naive_autoreg_baselines.make_predictions(
test_df, model, best_window)
elif method == 'exponential':
preds = exponential_modeling.exponential_fit(df[outcome].values,
mode=mode,
target_day=target_day)
df[output_key] = preds
#del test_df['predicted_deaths_exponential']
return df
elif method == 'linear':
preds = exponential_modeling.linear_fit(df[outcome].values,
mode=mode,
target_day=target_day)
df[output_key] = preds
#del test_df['predicted_deaths_exponential']
return df
elif method == 'shared_exponential':
# Fit a poisson GLM with shared parameters across counties. Input to the poisson GLM is demographic_vars and log(previous_days_deaths+1)
cur_day_predictions = exponential_modeling.fit_and_predict_shared_exponential(
df,
mode,
outcome=outcome,
demographic_vars=demographic_vars,
target_day=target_day)
#if len(demographic_vars) > 0:
# output_key += '_demographics'
# import IPython
# IPython.embed()
df[output_key] = cur_day_predictions
return df
elif method == 'ensemble':
print('please use fit_and_predict_ensemble instead')
else:
print('Unknown method')
raise ValueError
def fit_and_predict(df,
outcome: str = 'deaths',
method: str = 'exponential',
mode: str = 'predict_future',
target_day: np.ndarray = np.array([1]),
output_key: str = None,
demographic_vars=[],
verbose: bool = False):
"""
Trains a method (method) to predict a current number of days ahead (target_day)
Predicts the values of the number of deaths for the final day of test_df and writes to the column
'predicted_deaths_'+method+'_'+str(target_day[-1]) of the test_df
Params
------
df
a df with county level deaths and cases and demographic information
outcome
key for the outcome to predict (the values in this column should have a list for each row)
method
what method to use to do forecasting
target_day
np.array([1,2,..,n]) predicts these number of days ahead (can just be np.array([3])) for example if you just want 3 days ahead)
output_key
key to save the output as
mode:
either 'predict_future' or 'eval_mode'
predict_future is predicting deaths on FUTURE days, so target_day=np.array([1])) means it predicts tomorrow's deaths
eval_mode is for evaluating the performance of the classifier.
target_day=np.array([k])) will predict the current days death count using information from k days ago.
target_day= np.array([1,2,3,...,k]) will predict todays deaths, yesterdays deaths, deaths k-1 days ago using information from k days ago.
Returns
-------
test_df
returns dataframe with added column
"""
assert mode == 'predict_future' or mode == 'eval_mode', 'unknown mode'
if output_key is None:
output_key = f'predicted_{outcome}_{method}_{target_day[-1]}'
if len(demographic_vars) > 0:
output_key += '_demographics'
if method == 'AR':
print('currently deprecated')
raise NotImplementedError
loss, model, best_window = naive_autoreg_baselines.train_and_evaluate_model(
train_df, test_df)
return naive_autoreg_baselines.make_predictions(
test_df, model, best_window)
elif method == 'exponential':
preds = exponential_modeling.exponential_fit(df[outcome].values,
mode=mode,
target_day=target_day)
df[output_key] = preds
#del test_df['predicted_deaths_exponential']
return df
elif method == 'linear':
preds = exponential_modeling.linear_fit(df[outcome].values,
mode=mode,
target_day=target_day)
df[output_key] = preds
#del test_df['predicted_deaths_exponential']
return df
elif method == 'shared_exponential':
# Fit a poisson GLM with shared parameters across counties. Input to the poisson GLM is demographic_vars and log(previous_days_deaths+1)
cur_day_predictions = exponential_modeling.fit_and_predict_shared_exponential(
df,
mode,
outcome=outcome,
demographic_vars=demographic_vars,
target_day=target_day,
verbose=verbose)
#if len(demographic_vars) > 0:
# output_key += '_demographics'
# import IPython
# IPython.embed()
df[output_key] = cur_day_predictions
return df
elif method == 'ensemble':
print('please use fit_and_predict_ensemble instead')
elif method == 'advanced_shared_model':
if 'neighbor_deaths' not in df.columns:
neighboring_counties_df = pd.read_csv(
oj(
parentdir,
'data/county_level/raw/county_ids/county_adjacency2010.csv'
))
neighboring_counties_df['fipscounty'] = neighboring_counties_df[
'fipscounty'].astype(str).str.zfill(5)
neighboring_counties_df['fipsneighbor'] = neighboring_counties_df[
'fipsneighbor'].astype(str).str.zfill(5)
df['countyFIPS'] = df['countyFIPS'].astype(str).str.zfill(5)
county_neighbor_deaths = []
county_neighbor_cases = []
county_fips = list(df['countyFIPS'])
for fips in county_fips:
neighboring_counties = list(neighboring_counties_df.loc[
neighboring_counties_df['fipscounty'] == fips]
['fipsneighbor'])
neighboring_county_deaths = list(df.loc[df['countyFIPS'].isin(
neighboring_counties)]['deaths'])
neighboring_county_cases = list(df.loc[df['countyFIPS'].isin(
neighboring_counties)]['cases'])
# if not in county adjacency file, assume neighboring deaths/counts to 0
if len(neighboring_county_deaths) == 0:
n_deaths = len(
df.loc[df['countyFIPS'] == fips]['deaths'].iloc[0])
n_cases = len(
df.loc[df['countyFIPS'] == fips]['cases'].iloc[0])
sum_neighboring_county_deaths = np.zeros(n_deaths)
sum_neighboring_county_cases = np.zeros(n_cases)
else:
sum_neighboring_county_deaths = np.zeros(
len(neighboring_county_deaths[0]))
for deaths in neighboring_county_deaths:
sum_neighboring_county_deaths += deaths
sum_neighboring_county_cases = np.zeros(
len(neighboring_county_deaths[0]))
for cases in neighboring_county_cases:
sum_neighboring_county_cases += cases
county_neighbor_deaths.append(sum_neighboring_county_deaths)
county_neighbor_cases.append(sum_neighboring_county_cases)
df['neighbor_deaths'] = county_neighbor_deaths
df['neighbor_cases'] = county_neighbor_cases
feat_transforms = defaultdict(lambda y: [lambda x: x])
feat_transforms['deaths'] = [lambda x: np.log(x + 1)]
feat_transforms['cases'] = [lambda x: np.log(x + 1)]
feat_transforms['neighbor_deaths'] = [lambda x: np.log(x + 1)]
feat_transforms['neighbor_cases'] = [lambda x: np.log(x + 1)]
default_values = defaultdict(lambda: 0)
aux_feats = ['cases', 'neighbor_deaths', 'neighbor_cases']
shared_model = SharedModel(df=df,
outcome=outcome,
demographic_variables=[],
mode=mode,
target_days=target_day,
feat_transforms=feat_transforms,
auxiliary_time_features=aux_feats,
time_series_default_values=default_values,
scale=True)
shared_model.create_dataset()
shared_model.fit_model()
shared_model.predict()
df[output_key] = shared_model.predictions
return df
else:
print('Unknown method')
raise ValueError
def fit_and_predict(train_df,
test_df,
outcome,
method,
mode,
target_day=np.array([1]),
demographic_vars=[]):
"""
Trains a method (method) to predict a current number of days ahead (target_day)
Predicts the values of the number of deaths for the final day of test_df and writes to the column
'predicted_deaths_'+method+'_'+str(target_day[-1]) of the test_df
Input:
train_df, tests: dfs with county level deaths and cases
method: string
target_day = np.array([1,2,..,n]) predicts these number of days ahead (can just be np.array([3])) for example if you just want 3 days ahead)
mode: either 'predict_future' or 'eval_mode'
predict_future is predicting deaths on FUTURE days, so target_day=np.array([1])) means it predicts tomorrow's deaths
eval_mode is for evaluating the performance of the classifier. target_day=np.array([k])) will predict the current days death count
using information from k days ago. target_day= np.array([1,2,3,...,k]) will predict todays deaths, yesterdays deaths, deaths k-1 days ago
using information from k days ago.
Output:
test_df
"""
assert mode == 'predict_future' or mode == 'eval_mode', 'unknown mode'
if method == 'AR':
print('currently deprecated')
raise NotImplementedError
loss, model, best_window = naive_autoreg_baselines.train_and_evaluate_model(
train_df, test_df)
return naive_autoreg_baselines.make_predictions(
test_df, model, best_window)
elif method == 'exponential':
preds = exponential_modeling.exponential_fit(test_df[outcome].values,
mode=mode,
target_day=target_day)
test_df[f'predicted_{outcome}_{method}_{target_day[-1]}'] = preds
#del test_df['predicted_deaths_exponential']
return test_df
elif method == 'shared_exponential':
# Fit a poisson GLM with shared parameters across counties. Input to the poisson GLM is demographic_vars and log(previous_days_deaths+1)
cur_day_predictions = exponential_modeling.fit_and_predict_shared_exponential(
train_df,
test_df,
mode,
outcome=outcome,
demographic_vars=demographic_vars,
target_day=target_day)
save_name = f'predicted_{outcome}_{method}_{target_day[-1]}'
if len(demographic_vars) > 0:
save_name += '_demographics'
# import IPython
# IPython.embed()
test_df[save_name] = cur_day_predictions
return test_df
elif method == 'ensemble':
#if target_day != np.array([1]):
# raise NotImplementedError
shared_preds = exponential_modeling.fit_and_predict_shared_exponential(
train_df,
test_df,
mode=mode,
outcome=outcome,
demographic_vars=demographic_vars,
target_day=target_day)
exp_preds = exponential_modeling.exponential_fit(
test_df[outcome].values, mode=mode, target_day=target_day)
if mode == 'predict_future':
use_df = test_df
else:
use_df = exponential_modeling.leave_t_day_out(
test_df, target_day[-1])
weights = pmdl_weight.compute_pmdl_weight(
use_df,
methods=['exponential', 'shared_exponential'],
outcome=outcome)
weights_sum = weights['exponential'] + weights['shared_exponential']
preds = [
exp_preds[i] * weights['exponential'][i] / weights_sum[i] +
np.array(shared_preds[i]) * weights['shared_exponential'][i] /
weights_sum[i] for i in range(len(test_df))
]
test_df[f'predicted_{outcome}_{method}_{target_day[-1]}'] = preds
return test_df
else:
print('Unknown method')
raise ValueError