Fit renewal count processes regression models

Fit renewal regression models for count data via maximum likelihood.

Usage

renewalCount(
  formula,
  data,
  subset,
  na.action,
  weights,
  offset,
  dist = c("weibull", "weibullgam", "custom", "gamma", "gengamma"),
  anc = NULL,
  convPars = NULL,
  link = NULL,
  time = 1,
  control = renewal.control(...),
  customPars = NULL,
  seriesPars = NULL,
  weiMethod = NULL,
  computeHessian = TRUE,
  standardise = FALSE,
  standardise_scale = 1,
  model = TRUE,
  y = TRUE,
  x = FALSE,
  ...
)

Arguments

formula: a formula object. If it is a standard formula object, the left hand side specifies the response variable and the right hand sides specifies the regression equation for the first parameter of the conditional distribution. formula can also be used to specify the ancilliary regressions, using the operator `|`, see section ‘Details’.
data, subset, na.action,: arguments controlling formula processing via model.frame.
weights: optional numeric vector of weights.
offset: optional numeric vector with an a priori known component to be included in the linear predictor of the count model. Currently not used.
dist: character, built-in distribution to be used as the inter-arrival time distribution or "custom" for a user defined distribution, see section ‘Details’. Currently the built-in distributions are "weibull", "weibullgam", "gamma", "gengamma" (generalized-gamma) and "burr".
anc: a named list of formulas for ancillary regressions, if any, otherwise NULL. The formulas associated with the (exact) parameter names are used. The left-hand sides of the formulas in anc are ignored.
convPars: a list of convolution parameters arguments with slots nsteps, extrap and convMethod, see dCount_conv_bi. If NULL, default parameters will be applied.
link: named list of character strings specifying the name of the link functions to be used in the regression. If NULL, the canonical link function will be used, i.e, log if the parameter is supposed to be positive, identity otherwise.
time: numeric, time at which the count is observed; default to unity (1).
control: a list of control arguments specified via renewal.control.
customPars: list, user inputs if dist = "custom", see section ‘Details’.
seriesPars: list, series expansion input parameters with slots terms (number of terms in the series expansion), iter (number of iteration in the accelerated series expansion algorithm) and eps (tolerance in the accelerated series expansion algorithm), Only used if dist = "weibull" and weiMethod = c("series_mat", "series_acc").
weiMethod: character, computation method to be used if dist = "weibull" or "weibullgam", see dWeibullCount and dWeibullgammaCount.
computeHessian: logical, should the hessian (and hence the covariance matrix) be computed numerically at the fitted values.
standardise: logical, should the covariates be standardised using standardize::standardize() function.
standardise_scale: numeric the desired scale for the covariates; defaults to 1.
model, y, x: logicals. If TRUE the corresponding components of the fit (model frame, response, model matrix) are returned.
...: arguments passed to renewal.control in the default setup.

Value

An S3 object of class "renewal", which is a list with components including:

coefficients: values of the fitted coefficients.
residuals: vector of weighted residuals \(\omega * (observed - fitted)\).
fitted.values: vector of fitted means.
optim: data.frame output of optimx.
method: optimisation algorithm.
control: the control arguments, passed to optimx.
start: starting values, passed to optimx.
weights: weights to apply, if any.
n: number of observations (with weights > 0).
iterations: number of iterations in the optimisation algorithm.
execTime: duration of the optimisation.
loglik: log-likelihood of the fitted model.
df.residual: residuals' degrees of freedom for the fitted model.
vcoc: convariance matrix of all coefficients, computed numerically from the hessian at the fitted coefficients (if computeHessian is TRUE).
dist: name of the inter-arrival distribution.
link: list, inverse link function corresponding to each parameter in the inter-arrival distribution.
converged: logical, did the optimisation algorithm converge?
data: data used to fit the model.
formula: the original formula.
call: the original function call.
anc: named list of formulas to model regression on ancillary parameters.
score_fct: function to compute the vector of scores defined in Cameron and Trivedi (2013) , equation 2.94.
convPars: convolution inputs used.
customPars: named list, user passed distribution inputs, see section ‘Details’.
time: observed window used, default is 1.0 (see inputs).
model: the full model frame (if model = TRUE).
y: the response count vector (if y = TRUE).
x: the model matrix (if x = TRUE).

Details

renewal re-uses design and functionality of the basic R tools for fitting regression model (lm, glm) and is highly inspired by hurdle() and zeroinfl() from package pscl. Package Formula is used to handle formulas.

Argument formula is a formula object. In the simplest case its left-hand side (lhs) designates the response variable and the right-hand side the covariates for the first parameter of the distribution (as reported by getParNames. In this case, covariates for the ancilliary parameters are specified using argument anc.

The ancilliary regressions, can also be specified in argument formula by adding them to the righ-hand side, separated by the operator ‘|’. For example Y | shape ~ x + y | z can be used in place of the pair Y ~ x + y and anc = list(shape = ~z). In most cases, the name of the second parameter can be omitted, which for this example gives the equivalent Y ~ x + y | z. The actual rule is that if the parameter is missing from the left-hand side, it is inferred from the default parameter list of the distribution.

As another convenience, if the parameters are to to have the same covariates, it is not necessary to repeat the rhs. For example, Y | shape ~ x + y is equivalent to Y | shape ~ x + y | x + y. Note that this is applied only to parameters listed on the lhs, so Y ~ x + y specifies covariates only for the response variable and not any other parameters.

Distributions for inter-arrival times supported internally by this package can be chosen by setting argument "dist" to a suitable character string. Currently the built-in distributions are "weibull", "weibullgam", "gamma", "gengamma" (generalized-gamma) and "burr".

Users can also provide their own inter-arrival distribution. This is done by setting argument "dist" to "custom", specifying the initial values and giving argument customPars as a list with the following components:

parNames: character, the names of the parameters of the distribution. The location parameter should be the first one.
survivalFct: function object containing the survival function. It should have signature function(t, distPars) where t is the point where the survival function is evaluated and distPars is the list of the distribution parameters. It should return a double value.
extrapolFct: function object computing the extrapolation values (numeric of length 2) from the value of the distribution parameters (in distPars). It should have signature function(distPars) and return a numeric vector of length 2. Only required if the extrapolation is set to TRUE in convPars.

Some checks are done to validate customPars but it is user's responsibility to make sure the the functions have the appropriate signatures.

Note: The Weibull-gamma distribution is an experimental version and should be used with care! It is very sensitive to initial values and there is no guarantee of convergence. It has also been reparameterized in terms of \((1/r, 1/\alpha, c)\) instead of \((r, \alpha, c)\), where \(r\) and \(\alpha\) are the shape and scale of the gamma distribution and \(c\) is the shape of the Weibull distribution.

(2017-08-04(Georgi) experimental feature: probability residuals in component 'probResiduals'. I also added type 'prob' to the method for residuals() to extract them.

probResiduals[i] is currently 1 - Prob(Y[i] given the covariates). "one minus", so that values close to zero are "good". On its own this is probably not very useful but when comparing two models, if one of them has mostly smaller values than the other, there is some reason to claim that the former is superior. For example (see below), gamModel < poisModel in 3:1

References

Kharrat T, Boshnakov GN, McHale I, Baker R (2019). “Flexible Regression Models for Count Data Based on Renewal Processes: The Countr Package.” Journal of Statistical Software, 90(13), 1–35. doi:10.18637/jss.v090.i13 .

Cameron AC, Trivedi PK (2013). Regression analysis of count data, volume 53. Cambridge university press.

Examples

if (FALSE) { # \dontrun{
## may take some time to run depending on your CPU
data(football)
wei = renewalCount(formula = homeTeamGoals ~ 1,
                    data = football, dist = "weibull", weiMethod = "series_acc",
                    computeHessian = FALSE, control = renewal.control(trace = 0, 
                    method = "nlminb"))
} # }