Prophet

Quickstart

from sktime.forecasting.fbprophet import Prophet

estimator = Prophet(freq=None, add_seasonality=None, add_country_holidays=None, growth='linear', growth_floor=0.0, growth_cap=None, changepoints=None, n_changepoints=25, changepoint_range=0.8, yearly_seasonality='auto', weekly_seasonality='auto', daily_seasonality='auto', holidays=None, seasonality_mode='additive', seasonality_prior_scale=10.0, holidays_prior_scale=10.0, changepoint_prior_scale=0.05, mcmc_samples=0, alpha=0.05, uncertainty_samples=1000, stan_backend=None, verbose=0, fit_kwargs=None)

Parameters(22)

freq: str, default=None

A DatetimeIndex frequency. For possible values see https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html

add_seasonality: dict or None, default=None

Dict with args for Prophet.add_seasonality(). Dict can have the following keys/values:

• name: string name of the seasonality component.

• period: float number of days in one period.

• fourier_order: int number of Fourier components to use.

• prior_scale: optional float prior scale for this component.

• mode: optional ‘additive’ or ‘multiplicative’

• condition_name: string name of the seasonality condition.

add_country_holidays: dict or None, default=None

Dict with args for Prophet.add_country_holidays(). Dict can have the following keys/values:

• country_name: Name of the country, like ‘UnitedStates’ or ‘US’

growth: str, default=”linear”

String 'linear' or 'logistic' to specify a linear or logistic trend. If 'logistic' specified float for 'growth_cap' must be provided.

growth_floor: float, default=0

Growth saturation minimum value. Used only if growth="logistic", has no effect otherwise (if growth is not "logistic").

growth_cap: float, default=None

Growth saturation maximum aka carrying capacity. Mandatory (float) iff growth="logistic", has no effect and is optional, otherwise (if growth is not "logistic").

changepoints: list or None, default=None

List of dates at which to include potential changepoints. If not specified, potential changepoints are selected automatically.

n_changepoints: int, default=25

Number of potential changepoints to include. Not used if input changepoints is supplied. If changepoints is not supplied, then n_changepoints potential changepoints are selected uniformly from the first changepoint_range proportion of the history.

changepoint_range: float, default=0.8

Proportion of history in which trend changepoints will be estimated. Defaults to 0.8 for the first 80%. Not used if changepoints is specified.

yearly_seasonality: str or bool or int, default=”auto”

Fit yearly seasonality. Can be 'auto', True, False, or a number of Fourier terms to generate.

weekly_seasonality: str or bool or int, default=”auto”

Fit weekly seasonality. Can be 'auto', True, False, or a number of Fourier terms to generate.

daily_seasonality: str or bool or int, default=”auto”

Fit daily seasonality. Can be 'auto', True, False, or a number of Fourier terms to generate.

holidays: pd.DataFrame or None, default=None

pd.DataFrame with columns holiday (string) and ds (date type) and optionally columns lower_window and upper_window which specify a range of days around the date to be included as holidays. lower_window=-2 will include 2 days prior to the date as holidays. Also optionally can have a column prior_scale specifying the prior scale for that holiday.

seasonality_mode: str, default=’additive’

One of 'additive' or 'multiplicative'.

seasonality_prior_scale: float, default=10.0

Parameter modulating the strength of the seasonality model. Larger values allow the model to fit larger seasonal fluctuations, smaller values dampen the seasonality. Can be specified for individual seasonalities using add_seasonality.

holidays_prior_scale: float, default=10.0

Parameter modulating the strength of the holiday components model, unless overridden in the holidays input.

changepoint_prior_scale: float, default=0.05

Parameter modulating the flexibility of the automatic changepoint selection. Large values will allow many changepoints, small values will allow few changepoints.

mcmc_samples: int, default=0

If greater than 0, will do full Bayesian inference with the specified number of MCMC samples. If 0, will do MAP estimation.

alpha: float, default=0.05

Width of the uncertainty intervals provided for the forecast. If mcmc_samples=0, this will be only the uncertainty in the trend using the MAP estimate of the extrapolated generative model. If mcmc.samples>0, this will be integrated over all model parameters, which will include uncertainty in seasonality.

uncertainty_samples: int, default=1000

Number of simulated draws used to estimate uncertainty intervals. Settings this value to 0 or False will disable uncertainty estimation and speed up the calculation.

stan_backend: str or None, default=None

str as defined in StanBackendEnum. If None, will try to iterate over all available backends and find the working one.

fit_kwargs: dict or None, default=None

Dict with args for Prophet.fit(). These are additional arguments passed to the optimizing or sampling functions in Stan.

Examples

>>> from sktime.datasets import load_airline
>>> from sktime.forecasting.fbprophet import Prophet
>>> # Prophet requires to have data with a pandas.DatetimeIndex
>>> y = load_airline (). to_timestamp (freq = 'M')
>>> forecaster = Prophet (
... seasonality_mode = 'multiplicative',
... n_changepoints = int (len (y) / 12),
... add_country_holidays = { 'country_name': 'Germany' },
... yearly_seasonality = True)
>>> forecaster. fit (y) Prophet(
... )
>>> y_pred = forecaster. predict (fh = [1, 2, 3 ])

References