Per-dimension operator: list of ``(coef, u_value, differentiate)``. Applying the tensor product of the two dimensions' operators to ``C``

Hook to adjust the fitted ``NonParametricCounting`` model in place before it is returned. Default is a no-op.

Group the event times by system and resolve each system's observation window. Returns a list of ``(events_used, T_q)`` per system (with the

Convert ``Z`` to a numeric design matrix, selecting the covariate columns recorded at fit time when a pandas DataFrame is passed.

Simulate ``items`` count-terminated sequences and return the raw event data as an ``xicn`` dict (``events + 1`` exact events per sequ

Simulate ``items`` time-terminated sequences and return the raw event data as an ``xicn`` dict. Each sequence ends in a right-censore

Standardise ``x`` onto the unit interval.

Force the Turnbull heuristic when the data is interval censored or truncated, warning that the requested heuristic was changed.

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

Joint CDF ``P(X_1 <= x_1, X_2 <= x_2)``.

Coerce the ``x`` variable of xcnt-format data into a numpy array. Accepts a 1D array of event values, or a 2D array / list-of-pairs of `

Approximate parameter covariance matrix, ordered to match :attr:`parameter_names`. Computed as the inverse of the numerical H

Draw the probability plot described by the ``probability_plot_data`` dictionary ``d`` onto the matplotlib axes ``ax``.

``dC/dv``. autograd default; override if known.

Minimise ``fun`` trying Newton-CG with the supplied jacobian and hessian first, then falling back to BFGS, then Nelder-Mead, whenever a m

r""" The cumulative distribution function, or failure function, for a distribution using the parameters found in the ``.params`` attr

r""" Failure (CDF or unreliability) function for the LogNormal Distribution: .. math:: F(x) = \Phi \left( \frac{\ln(x) -

r""" Failure (CDF or unreliability) function for the Beta Distribution: .. math:: F(x) = \int_{0}^{x}t^{\alpha-1}\left (

r""" Failure (CDF or unreliability) function for the four-parameter Beta distribution: .. math:: F(x) = I_{z}\le

Fits Cox Proportional Hazards model to the provided data. Parameters ---------- x: array-like The obser

Fit the copula and its margins to multivariate survival data. Parameters ---------- x, c, n, t, xl, xr Multivaria

Fit the ARI model from recurrent data. Parameters ---------- data : RecurrentData Data containing the r

Build the ``(n_rows, 2)`` truncation array from either a ``t`` matrix or separate ``tl``/``tr`` bounds (scalars broadcast to all rows). The d

Instantaneous intensity function (IIF) or the failure rate of the HPP model. Parameters ---------- x : array

Data on the survival of patients with advanced lung cancer from [6]_. References ---------- .. [6] Loprinzi CL. Laurie JA. Wieand

r""" Returns the best feature index and value according to the Log-Rank split criterion. That is, it returns .. math:: (u^*

r""" Mean of the Uniform distribution .. math:: E = \frac{1}{2} \left ( a + b \right ) Parameters -----

r""" Mean of the Beta distribution .. math:: E = \frac{\alpha}{\alpha + \beta} Parameters ----------

r""" Mean of the Rayleigh distribution .. math:: E = \sigma \sqrt{\frac{\pi}{2}} Parameters ----------

r""" Two-sided confidence interval of the (restricted) mean survival time, using the normal approximation with the standard variance

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

Convert a covariate input ``Z`` into a numeric design matrix. If ``Z`` is a pandas DataFrame, the columns are selected using the ``featu

Compute everything needed to draw a probability plot of the data against the fitted CDF ``ff``. ``dist`` provides the plotting configura

r""" Two-sided confidence interval of the quantile at each probability p using the Brookmeyer-Crowley method: the interval is

The nonparametric MCF estimators (``NonParametricCounting`` and ``CauseSpecificMCF``) currently only support exact events (``c=0``) and r

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

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

Spearman's rho. Default: empirical estimate from a large sample.

Simulate time-terminated recurrence data and return the raw events. Unlike :meth:`time_terminated_simulation` (which returns the fit

r""" Cumulative hazard rate for the Gumbel LEV Distribution: .. math:: H(x) = -\ln \left ( R(x) \right ) Parame

Assemble the parameter vector and covariance used by ``cb``. The variance is computed over the extended parameter vector ``(*params,

Confidence bound on the hazard (``hf``) or density (``df``). Both are non-negative, so the bound is computed on the log scale to keep

Build a ``NonParametricCounting`` model from an explicit ``(x, r, d)`` triple so its ``mcf``/``mcf_cb``/``plot`` machinery can be reused for

First Debye function ``D_1(t) = (1/t) int_0^t s/(e^s-1) ds``.

Initial parameter guess. Override per family for robustness.

Initial parameters for the baseline intensity model: the plain NHPP fit of that baseline if it succeeds, otherwise its own parameter

Monte-Carlo KL(true || fit) in nats, using the offset log-density.

Names of the entries of ``_mle``, in order. Subclasses override this to label their parameters (e.g. the renewal models prepend the r