NEWS.md
nc_estimate_*
function output the full model list as an attribute, that is really only necessary for those interested in the underlying models used for classifying the effectsnc_estimate_*_links()
functions to set thresholds for classifying links (#157)as_edge_tbl()
(#142)nc_classify_effects()
and nc_filter_estimates()
, merged them into the two main estimation functions insteadlm
and glm
models were removed for improving computing speed (they slowed things down quite a bit)nc_estimate_*
functions output@seealso
lm
and glm
models, model summary statistics are added (#88).nc_plot_network()
(#89, #110).nc_adjacency_graph()
, nc_adjacency_matrix()
, and nc_partial_corr_matrix()
to help create the weights for the network plot. (Issue #80, PR #89).pcor()
(#125, #131).nc_filter_estimates()
(#109).nc_standardize()
that prevented the ability to use the .regressed_on
. argument to extract residuals (#108).nc_standardize()
function to standardize the metabolic variables (#73).matches()
or starts_with()
(#73).net_coupler_out()
, getExp.coef.permetabolite()
, and getExp.coef.out()
(#59)nc_exposure_estimates()
nc_outcome_estimates()
function. Because of this streamlining, the code is much faster and with the move to use MuMIn we can remove our dependency on rJava via glmulti.nc_create_network()
function so that only the graph skeleton is output (#55).nc_create_network()
.nc_create_network()
and the outcome estimation functions. Travis and code coverage were added as well.nc_make_network()
to nc_create_network()
and moved into own file.nc_make_network()
code and moved into another file.