PREDICT Class

class PREDICT.PREDICT(data, model, dateCol='date', startDate='min', endDate='max', timestep='month', recalPeriod=None, verbose=False)

A class used to represent the PREDICT model.

Parameters:

  • data (pd.DataFrame) – The data to be used by the model.

  • model (PREDICTModel) – The model to be used for prediction.

  • dateCol (str)

  • startDate (str, dt.datetime) – The start date for the prediction window.

  • endDate (str, pd.datetime) – The end date for the prediction window.

  • timestep (str, int) – The timestep for the prediction window. Must be ‘week’, ‘day’, ‘month’, or an integer representing the number of days. Defaults to ‘month’.

  • currentWindowStart (pd.datetime) – The current start date of the prediction window.

  • currentWindowEnd (pd.datetime) – The current end date of the prediction window.

  • log (dict) – A dictionary to store logs.

  • logHooks (list) – A list of hooks to be called during logging.

  • recalPeriod (int) – An integer giving the number of days for the recalibration window. Defaults to ‘timestep’.

  • verbose (bool) – Print sample size calculation warnings.

addLog(key, date, val)

Adds a log entry to the log dictionary.

Parameters:

  • key (str) – The key for the log entry.

  • date (any) – The date for the log entry.

  • val (any) – The value for the log entry.

addLogHook(hook)

Adds a hook to the logHooks list.

Parameters:

hook (function) – A function to be called during logging.

getLog()

Returns the log dictionary.

Returns:

The log dictionary.

Return type:

dict

run()

Runs the prediction model over the specified date range.