Question on eliminating repeated generated expressions.

[AaronRoseDA]

Hi,

I have been using your package and it's been blowing my mind. It's really powerful stuff.

I do have a question on the possibility of removing or blacklisting functions that have already been used OR disallowing creating expressions that will return NAs.

When I use the function GrammaticalEvolution it often generates identical expressions. I understand why this happens and that these functions are technically not identical. However, I'm wondering if there is a parameter I can tweak or if my grammarDef could be improved to prevent identical functions. Possibly disallow function combinations that results in an error (sqrt(-1), Log(-1), etc..) Also, in my eval function (SSE) if NAs are produced it returns Inf. This might also be a weakness so any recommendations would be helpful.

If you need anything else from me please let me know. I appreciate your time!

warnings()
Warning messages:
1: In log(log(data$y)) : NaNs produced
...
14: In log(log(data$y)) : NaNs produced
15: In log(log(data$y)) : NaNs produced
...
38: In log(log(data$y)) : NaNs produced
39: In log(log(data$y)) : NaNs produced
40: In log(log(data$y)) : NaNs produced
...

Here's my grammarDef from my wrapper function below:

normGrammarDef$grammarDef
::= (, ) | () |
::= log
::= / | ^
::= data$x | data$y | 2

Here's my example data definition. It's a simple function for demo purposes.:

x <- runif(10, 0, 100)
y <- runif(10, 0, 1)

data <- data.frame(
  distType = "generic function",
  x=x,
  y=y,
  entropy = log(y/x) ^ 2
)

Here's my grammar definition and wrapper function:

create_grammar_wrapper <- function(data,
                                   operators = c("+", "-", "*", "/", "^"),
                                   functions = c('log', 'sqrt', 'exp'),
                                   variables,
                                   constants = c(1, 2, pi)) {

  ivCol <- variables  # Preserve variable names
  variables <- syms(paste0("data$",ivCol))

  functions <- syms(functions)
  operators <- syms(operators)

  # Define grammar rules
  ruleDef <- list(
    expr = grule(op(expr, expr), func(expr), var),
    func = do.call(grule, functions),  # Ensure function set includes `exp`
    op = do.call(grule, operators),
    var = do.call(grule, c(variables, constants))
  )

  # Create the grammar definition
  grammarDef <- CreateGrammar(ruleDef)

  return(c(
    list(
      name = variables,
      known_entropy = data$entropy,
      grammarDef = grammarDef
    ),
    setNames(lapply(ivCol, function(col) data[[col]]), ivCol)  # Dynamically add multiple elements
  ))
}