seperate_decomp: get_log_decomp: calculate the decomp for all true & fake...
In drpmcain/decomp:

get_log_decomp: calculate the decomp for all true & fake models & sub-models by using the algorithm described in 'decomp_log'

1	seperate_decomp(is.output.final = T, model_name = NULL)

`model_name`	if NULL, then output the passed-thru model on the first layer from input_layer. Specific the model_name if need. I.E. model_name="web"
`input_data`	a data.table class dataset for modeling.
`input_var`	a data.table class dataset for variable coefficient.
`input_layer`	a data.table class dataset for model structure.
`decomp_list`	is from the result of get_decomp, it should be a list contains all the models' decomp
`layer`	is the set up file layer, which specified the structure of the model
`input_var`	a data.table class dataset for variable coefficient.
`cs`	a charactore vector for the cross section variable names in input_data. It could be more than one variable.
`date`	a charactor of the date variable name in input_data.
`is.output`	if TRUE, then output tabels to work directory. Default is False.

a list of modeling result. It contains the list of decomp for all models

get_reduced_model:

a list of modeling result. It contains the following components: contribution: contribution table; decomp: decomp table; for.plot: table for shiny; app: a shiny app;

drpmcain/decomp documentation built on May 15, 2019, 2:51 p.m.