R: intMarkovOrd

intMarkovOrd {Hmisc}

R Documentation

intMarkovOrd

Description

Compute Parameters for Proportional Odds Markov Model

Usage

intMarkovOrd(
  y,
  times,
  initial,
  absorb = NULL,
  intercepts,
  extra = NULL,
  g,
  target,
  t,
  ftarget = NULL,
  onlycrit = FALSE,
  constraints = NULL,
  printsop = FALSE,
  ...
)

Arguments

`y`	vector of possible y values in order (numeric, character, factor)
`times`	vector of measurement times
`initial`	initial value of `y` (baseline state; numeric, character, or factor matching `y`). If length 1 this value is used for all subjects, otherwise it is a vector of length `n`.
`absorb`	vector of absorbing states, a subset of `y` (numeric, character, or factor matching `y`). The default is no absorbing states. Observations are truncated when an absorbing state is simulated.
`intercepts`	vector of initial guesses for the intercepts
`extra`	an optional vector of intial guesses for other parameters passed to `g` such as regression coefficients for previous states and for general time trends. Name the elements of `extra` for more informative output.
`g`	a user-specified function of three or more arguments which in order are `yprev` - the value of `y` at the previous time, the current time `t`, the `gap` between the previous time and the current time, an optional (usually named) covariate vector `X`, and optional arguments such as a regression coefficient value to simulate from. The function needs to allow `yprev` to be a vector and `yprev` must not include any absorbing states. The `g` function returns the linear predictor for the proportional odds model aside from `intercepts`. The returned value must be a matrix with row names taken from `yprev`. If the model is a proportional odds model, the returned value must be one column. If it is a partial proportional odds model, the value must have one column for each distinct value of the response variable Y after the first one, with the levels of Y used as optional column names. So columns correspond to `intercepts`. The different columns are used for `y`-specific contributions to the linear predictor (aside from `intercepts`) for a partial or constrained partial proportional odds model. Parameters for partial proportional odds effects may be included in the ... arguments.
`target`	vector of target state occupancy probabilities at time `t`. If `extra` is specified, `target` must be a matrix where row names are character versions of `t` and columns represent occupancy probabilities corresponding to values of `y` at the time given in the row.
`t`	target times. Can have more than one element only if `extra` is given.
`ftarget`	an optional function defining constraints that relate to transition probabilities. The function returns a penalty which is a sum of absolute differences in probabilities from target probabilities over possibly multiple targets. The `ftarget` function must have two arguments: `intercepts` and `extra`.
`onlycrit`	set to `TRUE` to only return the achieved objective criterion and not print anything
`constraints`	a function of two arguments: the vector of current intercept values and the vector of `extra` parameters, returning `TRUE` if that vector meets the constrains and `FALSE` otherwise
`printsop`	set to `TRUE` to print solved-for state occupancy probabilities for groups 1 and 2 and log odds ratios corresponding to them
`...`	optional arguments to pass to `stats::nlm()`. If this is specified, the arguments that `intMarkovOrd` normally sends to `nlm` are not used.

Details

Given a vector intercepts of initial guesses at the intercepts in a Markov proportional odds model, and a vector extra if there are other parameters, solves for the intercepts and extra vectors that yields a set of occupancy probabilities at time t that equal, as closely as possible, a vector of target values.

Value

list containing two vectors named intercepts and extra unless oncrit=TRUE in which case the best achieved sum of absolute errors is returned

Author(s)

Frank Harrell