SAS standard score code generation Generic Function (EXPERIMENTAL)

EXPERIMENTAL STATE - MAY NOT WORK AS INTENDED

Score code will only be generated successfully for supported models. Other models and frameworks will be added in due time. Use create_scoreSample() to get a structure sample

Disclaimer: The score code that is generated is designed to be a working template for an R model, but is not guaranteed to work out of the box for scoring, publishing, or validating the model.

codegen(model, path, rds, libs, inputs, ...)

# S3 method for class 'lm'
codegen(
  model,
  path = "scoreCode.R",
  rds = "model.rds",
  libs = c(),
  inputs = NULL,
  ...
)

# S3 method for class 'glm'
codegen(
  model,
  path = "scoreCode.R",
  rds = "model.rds",
  libs = c(),
  inputs = NULL,
  cutoff = 0.5,
  ...
)

# S3 method for class 'workflow'
codegen(
  model,
  path = "scoreCode.R",
  rds = "model.rds",
  libs = c(),
  inputs = NULL,
  referenceLevel = NULL,
  ...
)

Arguments

model: model object (lm, glm, tidymodels workflow, ...)
path: file name and path to write
rds: .rds file name to be called
libs: vector of libraries to be added to the code. Some may be guessed from the type.
inputs: define inputs as the passed vector instead of guessed
...: to be passes to individual code generators
cutoff: classification probability cutoff
referenceLevel: reference level for a factor target value

Value

a code string

Methods (by class)

codegen(lm): Code generator for lm class models
codegen(glm): generator for glm class models, specifically logistic regression
codegen(workflow): generator for tidymodels workflow class models

Examples


if (FALSE) { # \dontrun{
# SAS viya doesn't play nice with variables with '.' in the names
colnames(iris) <- gsub("\\.", "_", colnames(iris))

# simple regression
model <- lm(Petal.Length ~ ., data = iris)

codegen(model)
} # }