remove `"auto"` and `"cor_auto"` from `dynEGA`
Closed this issue · 1 comments
AlexChristensen commented
Since dynEGA
computes the derivatives of the time series, all values (even categorical) will be continuous.
It is possible for there to be 7 or fewer continuous values resulting in "auto"
and "cor_auto"
to default to categorical correlations (polychoric/tetrachoric/polyserial/biserial). For "auto"
, a range error will be thrown in polychoric.matrix
if derivative values are less than 0
A workaround for now is to use corr = "pearson"
or corr = "spearman"
Once fixed, only "pearson"
and "spearman"
will be available by default
AlexChristensen commented
Resolution: treat continuous data with fewer categories than ordinal.categories
as categorical