Compared to version 0.3.5-4 of lphom, the new version:
adjust2integers()
function. In the previous version, instead of applying the solver
argument declared in the function, it incorrectly used the solver
argument declared when creating the x
object. It has also bounded the computational time required with argument solver = "lp_solve"
when performing integer linear programming. Now lphom requires package lpSolve
(>= 5.6.18).plot.lphom()
method. In the previous version, it produced an error when trying to plot an object generated with _dual
functions.Compared to version 0.3.1-1 of lphom, the new version:
apriori
and lambda
) in functions lphom()
, tslphom()
, nslphom()
and lclphom()
. The apriori
argument allows to introduce in the inference process information from polls or some other a priori knowledge/intuition about the row standardized transfer probabilities. The lambda
argument controls the weight the user assigns to the a priori information.rslphom()
function."ordinary"
, "enriched"
, "semifull"
, "fullreverse"
, "adjust1"
and "adjust2"
.lclphom()
algorithm by including the two possibilities of calculating the distance with the global matrix. The distance with the global matrix that is used to generate the local solution and the distance with the global matrix where the local solution integrates into. This is controlled with the type.errors
argument.lphom()
, tslphom()
, nslphom()
, lclphom()
, rslphom()
, lphom_joint()
,
tslphom_joint()
and nslphom_joint()
the list deterministic.bounds
. The object deterministic.bounds
contains both the deterministic global and unit bounds that derive from the observed margins.