Provides a general, flexible framework for estimating parameters
and empirical sandwich variance estimator from a set of unbiased estimating
equations (i.e., M-estimation in the vein of Stefanski & Boos (2002)

- renames the
`geex-summary`

class to`geex_summary`

- adds
`ee_i`

(observed estimating function values) to the component object - various other bug fixes

- Added
`grab_design_levels`

to find all factor-levels in a design matrix. This output can then be passed to the`xlev`

arg in`model.matrix`

or more importantly through`grab_psiFun`

for the models. This fixes issue 41.- Notably, if this seems to work well, then perhaps
`grab_design_levels`

could be added to the body of`grab_psiFun`

to automatically handle this issue.

- Notably, if this seems to work well, then perhaps
- Edited
`binomial_integrand`

to hopefully handle singleton groups better. It will now attempt to coerce back to a one-row matrix before calling`apply`

.

- fixes issue when using
`grab_psiFUN`

for`glm`

objects and non-grouped data. - removes option to use weights with
`grab_psiFUN`

, as it was not clear that was implemented correctly.

- adds basic examples for most functions, though to be sure, the vignettes provide more useful examples.

- requires R >= 3.3

- add a
`call`

slot to the S4`geex`

object. Now the`update`

function can be used to update elements of an`m_estimate`

call.

- implements an
`S4`

system throughout`geex`

`estimate_equations`

becomes`m_estimate`

. See documentation for changes to arguments. Notably,`eeFUN`

becomes`estFUN`

`make_eeFUN`

functions become`grab_psiFUN`

- speeds up the summation of list of matrices with
`compute_sum_of_list`

- plus many more updates and vignettes

- adds a
`weights`

argument to`estimate_equations`

for faster computations with grouped data. See the weights vignette for a demonstration.

- changes names of list items used in a
`corrections_list`

in`estimate_equations`

. Each item of the`correction_list`

must itself be a list at least one item:`correctFUN`

. Additional arguments to`correctFUN`

may be passed via`correctFUN_control`

.

- adds a vignette explaining how to use a different root finding algorithm via
`rootFUN`

argument in`estimate_equations`

- fixes issues where different
`rootFUN`

s would not work:`roots`

argument no longer needs to be set if`compute_roots = TRUE`

. Instead, starting values are passed via the`rootFUN_control`

list.- Removes the
`start`

argument from`compute_eeroots`

; set this option in`rootFUN_control`

.

- overhauls which arguments are passed to
`estimate_equations`

and how these arguments are parsed internally. See this function's documentation for details - adds
`geexex`

dataset for use in examples

- the inital
`geex`

release