These functions perform the updating step of s_vec
and w_vec
in
the ADMM algorithm for different types of penalty functions.
The updating steps of s_vec
and w_vec
share the same pattern in
the ADMM algorithm. These functions correspond to different penalty types
(p1_type
and p2_type
) provided from the original solver function.
UpdateSW_Identity
: identity update, this is needed when there
is no shrinkage applied, possibly since the corresponding lam
is
set to 0.
UpdateSW_Lasso
: Lasso update
UpdateSW_SCAD
: SCAD update
UpdateSW_MCP
: MCP update
RSAVS_UpdateSW_Identity(invec, param, const_r, const_bc)
RSAVS_UpdateSW_Lasso(invec, param, const_r, const_bc)
RSAVS_UpdateSW_SCAD(invec, param, const_r, const_bc)
RSAVS_UpdateSW_MCP(invec, param, const_r, const_bc)
numerical vector, the orignal s_vec
or w_vec
.
numerical vector, parameters needed for the penalty function.
numerical scalar, parameter needed in the quadratic term in the augmented part.
numerical scalar, parameter represents the weight of corresponding penalty part in the objective function
Please refer to the vignette about the algorithm detail design to find out more about how these parameters are defined.