starting.vals.PPO: Starting values from PPO

In metaSDTreg: Regression Models for Meta Signal Detection Theory

Starting values from PPO

Description

Used by metaSDTreg to calculate starting values using the partial proportional odds model as implemented in ordinal. Internal function.

Usage

starting.vals.PPO(formula0, data0, control0, L0, K0, n.covar0)

Arguments

formula0	See metaSDTreg.
data0	Data frame.
control0	See metaSDTcontrol.
L0	Number of levels in ordinal type 2 response.
K0	Number of levels of ordinal response 'A'.
n.covar0	Number of covariates aside from the 'signal' variable.

Details

The function calls the clm function using all.vars(update(formula0, . ~ . - signal)) as the formula argument and using ~ signal as the nominal argument (also see metaSDTreg for the meaning of the signal variable). The coefficients of the signal variable are weighted together within the noise- and signal-specific type 2 responses using their inverse standard errors as weights.

Value

A named vector of starting values.

metaSDTreg documentation built on March 31, 2023, 10:09 p.m.