mspec_new: Create, edit, and combine model specifications

In bcjaeger/rpriori: A grammar for testing apriori hypotheses in R

Description

mspec_empty(), mspec_add(), mspec_rmv(), and mspec_sub() are functions that can be used to create model fitting specifications. mspec_describe() is a function that helps clarify model specifications and facilitates interactive work.

Usage

mspec_new(name = "Model 1", ...)

mspec_empty(name = "Model 0")

mspec_add(mspec, ..., name = NULL)

mspec_rmv(mspec, name, ...)

mspec_sub(mspec, name, ...)

mspec_describe(mspec, verbose = FALSE)

Arguments

name	(character value). The label that will be used to describe this model specification
...	A set of variable names (separated by commas) that will be used to adjust estimates in fitted models when this specification is applied. The values may be supplied with or without quotation marks around the variable names (e.g., 'x' or x).
parent	This parameters if only used for mspec_add, mspec_rmv, and mspec_sub. The parent mspec_apri object passes it's control variables to the descendant, accounting for modifications specified in ...

Details

Use mspec_empty() to create a generic unadjusted specification. Use mspec_add(), mspec_rmv(), and mspec_sub() to create descendant model specifications, and use mspec_describe() to summarize the structure of the model specifications in a concise manner that accounts for their dependencies (see examples).

Value

• mspec_empty() returns an unadjusted model specification.

• mspec_new() returns a model specification

• mspec_add(), mspec_rmv() returns a model specification with the designated control variables added or removed from the parent model specification.

• mspec_sub() returns a model specification with the designated control variables (left hand side of the =) replaced by variable names supplied by the user (right hand side of the =)

Examples

# Make an unadjusted model
m0  <- mspec_empty("Model 0")

# mspec_describe(mspec) is the same as print(mspec)
mspec_describe(m0)
print(m0)

# Model 1 includes adjustment for sex and class
m1  <- mspec_add(m0, name = "Model 1", sex, pclass)

m1

# Model 2a = model 1 + no. of siblings/spouses
m2a <- mspec_add(m1, name = 'Model 2a', sibsp)

m2a

# Model 2b = model 1 + no. of parents/children
m2b <- mspec_add(m1, name = 'Model 2b', parch)

m2b

# Model 3 = model 1, swapping out class for ticket fare

m3 <- mspec_sub(m1, name = 'Model 3', pclass = fare)

m3

bcjaeger/rpriori documentation built on Nov. 4, 2019, 6:52 a.m.