degrees-of-freedom-helpers: df calculation for RM-DIF-model with penalization
In submission-user/cmlDIFlasso: DIF-detection with Lasso-methods in a cML-Framework

Description Usage Arguments Details Value

calculates the first and second parts of df_delta depending on DIF-type given parameter estimates for delta under a certain lambda and the unrestricted DIF-model.

df_pen_unequal_zero(deltas, DIF_type, tolerance = 0)

df_veclength(deltas, DIF_type, delta_unres)

df_lambda(deltas, DIF_type, delta_unres, tolerance)

`deltas`	values for DIF-parameters as matrix for certain Lambda per item (columns) per covariate (rows)
`DIF_type`	for which Lasso-term should the non-nil values be calculated? options: "items", "variables", "all.interactions"
`tolerance`	Value under which parameters are considered equal to zero for the calculation of degreees of freedom
`delta_unres`	values for DIF-parameters as matrix form unrestricted cml solution (lambda = 0)

Title

calculated first (df_pen_unequal_zero) and second part (df_veclength) of df_delta, or already the complete degrees of freedom (df_lambda) for a lasso penalized (cml) model according to Yuan and Lin(2006) and Zou et al. (2007); remark: 10^(-6) will be zero (default value of argument tolerance)

submission-user/cmlDIFlasso documentation built on Dec. 23, 2021, 6:40 a.m.