Create SurpyvalData instance from JSON string or file path. Parameters ---------- source : str | Path Pa

Converts the fsli format to the xcn format. This ensures is so that the data can be passed to one of the parametric or nonparametric fitters.

Per-item integration bounds for the NHPP likelihood's right window-close. The NHPP integral runs from each item's entry time

Data on the recidivism of released prisoners from [8]_. Uses only static covariates. References ---------- .. [8] Rossi, P.H.,

r""" The n-th moment of the distribution using the parameters found in the ``.params`` attribute. Parameters -------

Negative log-likelihood for a covariate-aware survival model. Parameters ---------- model : object Must provide ``log_df``, ``log

Virtual-age and history-dependent models (Kijima/G1/ARA/ARI) cannot be fitted to left-truncated (delayed-entry) data: the virtual age or

r""" Surival (or Reliability) function with the non-parametric estimates from the data. Parameters ----------

The survival function of the fitted model. Parameters ---------- x : array like The values at which the

r""" Survival (or reliability) function for the Logistic Distribution: .. math:: R(x) = 1 - \\frac{1}{1 + e^{- \\left (

r""" Survival (or Reliability) function for the LogNormal Distribution: .. math:: R(x) = 1 - \Phi \left( \frac{\ln(x) -

r""" Survival (or Reliability) function for the Uniform Distribution: .. math:: R(x) = \frac{b - x}{b - a} Para

r""" Survival (or reliability) function for the Beta Distribution: .. math:: R(x) = 1 - \int_{0}^{x}t^{\alpha-1}\left (1

r""" Survival (or reliability) function for the LogLogistic Distribution: .. math:: R(x) = 1 - \frac{1}{1 + \left ( x /

r""" Survival (or Reliability) function for the Gamma Distribution: .. math:: R(x) = 1 - \frac{\gamma \left ( \alpha, \b

r""" Survival (or Reliability) function for the Normal Distribution: .. math:: R(x) = 1 - \Phi \left( \frac{x - \mu}{\si

r""" Survival (reliability) function for the Binomial distribution: .. math:: R(x) = P(X > x) = 1 - F(x) Parame

r""" Survival (or reliability) function for the four-parameter Beta distribution: .. math:: R(x) = 1 - I_{z}\lef

r""" Survival (or reliability) function for the ExpoWeibull Distribution: .. math:: R(x) = 1 - \left [ 1 - e^{-\left ( \

r""" Survival (or reliability) function for the Gumbel LEV Distribution: .. math:: R(x) = 1 - e^{-e^{-\left ( x - \mu \r

r""" Kernel smoothed estimate of the hazard rate, using an Epanechnikov kernel over the increments of the cumulative hazard es

Convert the recurrent event data to xrd format. Parameters ---------- estimator : str, optional The esti

Convert input to float array with better error handling.

The Arithmetic Reduction of Age/Intensity models (``ARA``/``ARI``) are parameterised by an integer memory ``m`` (how many prior failures the

The renewal models only define likelihood contributions for exact events (``c=0``) and right-censored observations (``c=1``). Interval (``c=2

Create an Accelerated Life fitter for the given distribution and life model. Parameters ---------- distribution : ParametricFitt

r""" Cumulative hazard rate with the non-parametric estimates from the data. This is calculated using the relationship betwee

r""" Cumulative hazard rate for the Gumbel Distribution: .. math:: H(x) = e^{\frac{x-\mu}{\sigma}} Parameters

Build the (bounded -> unbounded) parameter transforms used by the fitters that optimise in an unconstrained space. ``bounds`` are the

Survival function including the LFP and zero-inflation mass.

Confidence bound on the survival function via a logit transform. Working on the logit of R keeps the bound within ``(0, 1)``. The ret

5th-order central diff of f(a, x) w.r.t. a. Both a and x must be plain numpy values — call with getval().

Exact floating-point step size for central differences.

Fit with an offset and return parameter vs distribution metrics.

r""" The Aikake Information Criterion (AIC) for the model, if it was fit with the ``fit()`` method. Not available if fit with the

Effective (virtual) age at the start of each interarrival for the Arithmetic Reduction of Age model with memory ``m`` (Doyen & Gaudoin, 2

Node factory. Decides to return IntermediateNode object, or its sibling TerminalNode.

Compute the cumulative incidence function (CIF) based on the fitted model. No need to pass parameters as it uses the parameters of th

r""" The conditional survival of the model; that is, the probability that an item that has survived to ``X`` survives a further ``x``

r""" Density function for the ExpoWeibull Distribution: .. math:: f(x) = \mu \left ( \frac{\beta}{\alpha} \right ) \left

Finds the interarrival times between events for each item. The class assumes that the time of the event is cumulative, sometimes it i

Compute the instantaneous incidence function of the model with the parameters found by the fit method. Parameters -

Natural logarithm of the instantaneous intensity function (IIF) of the HPP model. Parameters ---------- x :

r""" The median survival time; the smallest observed value at which the estimated CDF reaches, or exceeds, 0.5. NaN if the estimate

r""" The negative log-likelihood for the model, if it was fit with the ``fit()`` method. Not available if fit with the ``from_params(

r""" Survival (or Reliability) function for a distribution using the parameters found in the ``.params`` attribute. Paramete

r""" Survival (or reliability) function for the Rayleigh Distribution: .. math:: R(x) = e^{-\frac{x^2}{2\sigma^2}}

Lower/upper tail-dependence coefficients ``(lambda_L, lambda_U)``. Default ``(0.0, 0.0)`` (no tail dependence); families override.

Convert the recurrent event data to xrd format for a single event type (cause). The at-risk set ``r`` is shared across all causes (an

r""" Cumulative hazard rate for the Beta distribution. .. math:: H(x) = -\ln\left( R(x) \right ) Parameters

First-order delta-method variance: ``Var(g) = J Sigma J^T``.

Split an extended parameter vector into ``(core, p, f0)``.

5th-order central diff of f(a, b, x) w.r.t. a — all args plain numpy.

5th-order central diff of f(a, b, x) w.r.t. b — all args plain numpy.

The G1 renewal process scales interarrival times by ``(1 + q) ** j``. For the scaled times to remain valid the base distribution must

``model.ff(x, Z, *params)`` with non-finite bounds replaced by their limiting probability without ever evaluating ``ff`` at a non-finite value

Initial parameters for the underlying lifetime distribution, fitted to the times-to-first-event when there are enough of them (these

Numerically integrate ∫ xⁿ f(x) dx over the distribution's support.

MOM: Method of Moments for the beta distribution has an analytic answer

Return a callable ``sample(ui) -> xi`` that draws the next interarrival time from a uniform random number, maintaining any per-sequen