intermittent_forecast

hana_ml.algorithms.pal.tsa.intermittent_forecast.intermittent_forecast(data, key=None, endog=None, p=None, q=None, forecast_num=None, optimizer=None, method=None, grid_size=None, optimize_step=None, accuracy_measure=None, ignore_zero=None, expost_flag=None, thread_ratio=None, iter_count=None, random_state=None, penalty=None)

Intermittent Time Series Forecast (ITSF) is a forecasting strategy for products with intermittent demand.

Differences from the constant weight of the Croston method:

  • ITSF provides an exponential weight for estimation, meaning closer data points have greater weight.

  • ITSF does not require the initial value of non-zero demands or the time intervals between non-zero demands.

Parameters:
dataDataFrame

Input data.

keystr, optional

Specifies the ID (representing time-order) column of data.

Required if a single ID column cannot be inferred from the index of data.

If there is a single column name in the index of data, then key defaults to that column; otherwise, key is mandatory.

endogstr, optional

Specifies the name of the column for intermittent demand values.

Defaults to the first non-key column of data.

pint, optional

The smoothing parameter for demand, where:

  • -1 : Automatically optimizes this parameter.

  • Positive integers : Specifies the value for smoothing ([1, n]) and forecasts manually.

The specified value cannot exceed the length of the time series for analysis.

Defaults to -1.

qint, optional

The smoothing parameter for the time intervals between intermittent demands, where:

  • -1 : Automatically optimizes this parameter.

  • Non-negative values ([1, p]) : Specifies the value manually.

Defaults to -1.

forecast_numint, optional

Forecast length. When set to 1, the algorithm forecasts only one value.

Defaults to 1.

optimizer{'lbfgsb', 'brute', 'sim_annealing'}, optional

Specifies the optimization algorithm for automatically identifying parameters p and q.

  • 'lbfgsb' : Bounded Limited-memory Broyden-Fletcher-Goldfarb-Shanno (LBFGSB) method with parameters p and q initialized by the default scheme.

  • 'brute' : Brute-force method, LBFGSB with parameters p and q initialized by grid search.

  • 'sim_annealing' : Simulated annealing method.

Defaults to 'lbfgsb'.

methodstr, optional

Specifies the method (or mode) for the output:

  • 'sporadic': Uses the sporadic method.

  • 'constant': Uses the constant method.

Defaults to 'constant'.

grid_sizeint, optional

Specifies the number of steps from the start point to the length of the data for grid search.

Only valid when optimizer is set to 'brute'.

Defaults to 20.

optimize_stepfloat, optional

Specifies the minimum step for each iteration of the LBFGSB method.

Defaults to 0.001.

accuracy_measurestr or list of str, optional

The metric to quantify how well a model fits the input data. Options: 'mse', 'rmse', 'mae', 'mape', 'smape', 'mase'.

Defaults to 'mse'.

Note

Specify a measure name if you want the corresponding measure value to be reflected in the output statistics (the second DataFrame in the return).

ignore_zerobool, optional

  • False: Uses zero values in the input dataset when calculating 'mape'.

  • True: Ignores zero values in the input dataset when calculating 'mape'.

Only valid when accuracy_measure is 'mape'.

Defaults to False.

expost_flagbool, optional

  • False: Does not output the expost forecast and only outputs the forecast values.

  • True: Outputs both the expost forecast and the forecast values.

Defaults to True.

thread_ratiofloat, optional

Adjusts the percentage of available threads to use, from 0 to 1. A value of 0 indicates the use of a single thread, while 1 implies the use of all possible threads. Values outside the range will be ignored, and this function heuristically determines the number of threads to use.

Defaults to 0.

iter_countint, optional

A positive integer that controls the number of iterations for simulated annealing.

Defaults to 1000.

random_stateint, optional

Specifies the seed for the random number generator. Valid for the simulated annealing method.

Defaults to 1.

penaltyfloat, optional

A penalty applied to the cost function to avoid overfitting.

Defaults to 1.0.

Returns:
tuple of DataFrames

  • 1st DataFrame: Forecast values.

  • 2nd DataFrame: Related statistics.

Examples

>>> forecasts, stats = intermittent_forecast(data=df, p=3, forecast_num=3,
                                             optimizer='lbfgsb_grid', grid_size=20,
                                             optimize_step=0.011, expost_flag=False,
                                             accuracy_measure='mse', ignore_zero=False,
                                             thread_ratio=0.5)
>>> forecasts.collect()
>>> stats.collect()