Created
June 6, 2022 00:31
-
-
Save hassanj47/b808429a3c62f8494faf45b9a0626afa to your computer and use it in GitHub Desktop.
prophet model for time series survival analysis
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
def build_prophet(df): | |
'''Build prophet model and generate events for survival modelling.''' | |
# prepping data for prophet assuming data starts from Jan, 2015 | |
df['ds'] = pd.date_range(start='2015-01-01', periods=len(df)) | |
df['y'] = df.sales | |
df = df[['ds','y']] | |
# train/test splits | |
train = df[:-TEST_DAYS] | |
test = df[-TEST_DAYS:] | |
model = Prophet(daily_seasonality=True) | |
model.fit(train) | |
future = model.make_future_dataframe(FORECAST_DAYS, include_history=False) | |
forecast = model.predict(future) | |
results = pd.merge(forecast, test, on='ds') | |
# resample on weeks | |
results = results.set_index('ds').resample('w').sum().reset_index() | |
# generate metric suitable for survival analysis | |
results_survival = survival_mape_exp(results, WIN_SIZE_WEEKS) | |
# generate events and censors | |
thresh_week, event = survival_events(results_survival) | |
return pd.DataFrame({'thresh_week':[thresh_week], 'event':[event]}) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment