pollinate_ML {emery}

Generate seed values for EM algorithm

Description

pollinate_ML() is a general helper function which can be used to generate starting values, i.e. seeds, for the estimate_ML() function from a multi-method data set.

Usage

pollinate_ML(type = c("binary", "ordinal", "continuous"), data, ...)

pollinate_ML_binary(data, ...)

pollinate_ML_ordinal(
  data,
  n_level = NULL,
  threshold_level = ceiling(n_level/2),
  level_names = NULL
)

pollinate_ML_continuous(
  data,
  prev = 0.5,
  q_seeds = c((1 - prev)/2, 1 - (prev/2)),
  high_pos = TRUE
)

Arguments

type

A string specifying the data type of the methods under evaluation.

data

An n_obs by n_method matrix containing the observed values for each method. If the dimensions are named, row names will be used to name each observation (obs_names) and column names will be used to name each measurement method (method_names).

...

Additional arguments

n_level

Used for ordinal methods. Integer number of levels each method contains

threshold_level

Used for ordinal methods. A value from 1 to n_level which indicates the initial threshold used to define positive and negative disease states.

level_names

Used for ordinal methods. Optional vector of length n_level containing names for each level.

prev

A double between 0-1 representing the proportion of positives in the population

q_seeds

Used for continuous methods. A vector of length 2 representing the quantiles at which the two groups are assumed to be centered

high_pos

Used for continuous methods. A logical indicating whether larger values are considered "positive"

Value

a list of EM algorithm initialization values