pmdarima.arima.ndiffs

pmdarima.arima.ndiffs(x, alpha=0.05, test='kpss', max_d=2, **kwargs)

Estimate ARIMA differencing term, d.

Perform a test of stationarity for different levels of d to estimate the number of differences required to make a given time series stationary. Will select the maximum value of d for which the time series is judged stationary by the statistical test.

Parameters:

x : array-like, shape=(n_samples, [n_features])

The array (time series) to difference.

alpha : float, optional (default=0.05)

Level of the test. This is the value above below which the P-value will be deemed significant.

test : str, optional (default=’kpss’)

Type of unit root test of stationarity to use in order to test the stationarity of the time-series. One of (‘kpss’, ‘adf’, ‘pp’)

max_d : int, optional (default=2)

Maximum number of non-seasonal differences allowed. Must be a positive integer. The estimated value of d will not exceed max_d.

Returns:

d : int

The estimated differencing term. This is the maximum value of d such that d <= max_d and the time series is judged stationary. If the time series is constant, will return 0.

References

[R68]

R’s auto_arima ndiffs function https://github.com/robjhyndman/forecast/blob/19b0711e554524bf6435b7524517715658c07699/R/arima.R#L132 # noqa: E501

Examples using pmdarima.arima.ndiffs

Modeling quasi-seasonal trends with date features

Modeling quasi-seasonal trends with date features