NonHomogeneousPoissonProcess#

class relife.stochastic_process.NonHomogeneousPoissonProcess(lifetime_model)[source]#

Non-homogeneous Poisson process.

Methods

`cumulative_intensity`	The cumulative intensity function of the process.
`fit`	Estimation of the process parameters from recurrent failure data.
`freeze`	Freeze any arguments required by the process into the object data.
`generate_failure_data`	Generate failure data
`intensity`	The intensity function of the process.
`sample`	Renewal data sampling.

cumulative_intensity(time, *args)[source]#

The cumulative intensity function of the process.

Parameters:

timefloat or np.ndarray: Elapsed time value(s) at which to compute the function. If ndarray, allowed shapes are (), (n,) or (m, n).
*argsfloat or np.ndarray: Additional arguments needed by the model.

Returns:

np.float64 or np.ndarray: Function values at each given time(s).

fit(ages_at_events, events_assets_ids, assets_ids=None, first_ages=None, last_ages=None, lifetime_model_args=None, **options)[source]#

Estimation of the process parameters from recurrent failure data.

Parameters:

ages_at_events1D array of float: Array of float containing the ages of each asset when the events occured
events_assets_idssequence of hashable: Sequence object containing the ids of each assets corresponding to each ages.
first_ages1D array of float, optional: Array of float containing the ages of each asset before observing events. If set, assets_ids is needed and its length must equal the size of first_ages.
last_ages1D array of float, optional: Array of float containing the ages of each asset at the end of the observation period. If set, assets_ids is needed and its length must equal the size of last_ages.
lifetime_model_argstuple of np.ndarray, optional: Additional arguments needed by the model. If set, assets_ids is needed. For 1D array, the size must equal the length of assets_ids. For 2D array (e.g. covar of regression), the length of first axis must equal the length of assets_ids.
assets_idssequence of hashable, optional: Only needed if either first_ages, last_ages or model_args is filled. It must be a sequence object containing the unique ids corresponding to each values contained in first_ages, last_ages and/or model_args

Returns:

Self: The current object with the estimated parameters setted inplace.

Examples

Ages of assets AB2 and CX13 at each event.

>>> from relife.lifetime_model import Weibull
>>> from relife.stochastic_process import NonHomogeneousPoissonProcess
>>> nhpp = NonHomogeneousPoissonProcess(Weibull())
>>> nhpp.fit(
    np.array([11., 13., 21., 25., 27.]),
    ("AB2", "CX13", "AB2", "AB2", "CX13"),
)

With additional information and model args (regression of 2 coefficients)

>>> from relife.lifetime_model import ProportionalHazard
>>> nhpp = NonHomogeneousPoissonProcess(ProportionalHazard())
>>> nhpp.fit(
    np.array([11., 13., 21., 25., 27.]),
    ("AB2", "CX13", "AB2", "AB2", "CX13"),
    first_ages = np.array([10., 12.]),
    last_ages = np.array([35., 60.]),
    model_args = (np.array([[1.2, 5.5], [37.2, 22.2]]),) # 2d array of 2 raws (2 assets) and 2 columns (2 coefficients)
)

freeze(*args)[source]#

Freeze any arguments required by the process into the object data.

Parameters:

*argsfloat or np.ndarray: Additional arguments needed by the model.

Returns:

FrozenParametricModel

generate_failure_data(size, time_window, *args, seed=None)[source]#

Generate failure data

This function will generate failure data that can be used to fit a non-homogeneous Poisson process.

Parameters:

sizeint: The size of the desired sample
time_windowtuple of two floats: Time window in which failure data are sampled
seedint, optional: Random seed, by default None.

Returns:

A dict of ages_at_events, events_assets_ids, first_ages, last_ages, model_args and assets_ids

intensity(time, *args)[source]#

The intensity function of the process.

Parameters:

timefloat or np.ndarray: Elapsed time value(s) at which to compute the function. If ndarray, allowed shapes are (), (n,) or (m, n).
*argsfloat or np.ndarray: Additional arguments needed by the model.

Returns:

np.float64 or np.ndarray: Function values at each given time(s).

property nb_params#

Number of parameters.

Returns:

int: Number of parameters.

property params#

Parameters values.

Returns:

ndarray: Parameters values

Notes

If parameter values are not set, they are encoded as np.nan value.

property params_names#

Parameters names.

Returns:

list of str: Parameters names

Notes

Parameters values can be requested (a.k.a. get) by their name at instance level.

sample(nb_samples, time_window, *args, seed=None)[source]#

Renewal data sampling.

This function will sample data and encapsulate them in an object.

Parameters:

nb_samplesint: The number of samples
time_windowtuple of two floats: Time window in which data are sampled
*argsfloat or np.ndarray: Additional arguments needed by the model.
seedint, optional: Random seed, by default None.

NonHomogeneousPoissonProcess#

This Page