# R/minimalize.R In cna: Causal Modeling with Coincidence Analysis

#### Documented in minimalize

```# build dnf
#  Note that ct is supposed to be a "full" ct
make.dnf <- function(expr, ct, disj = "+"){
if (expr %in% c("0", "1")) return(expr)
selC <- selectCases(expr, ct)
if (nrow(selC) == 0L) return("0")
if (nrow(selC) == nrow(ct)) return("1")
sc <- ctInfo(selC)\$scores
terms <- split(rep(colnames(sc), each = nrow(sc))[sc == 1],
row(sc)[sc == 1])
conj <- C_mconcat(terms, "*")
conj <- conj[is.minimal(conj)]
C_concat(conj, disj)
}
# Minimize a single condition
#  Note that x is supposed to be a "full" ct
.minim1 <- function(cond, x, maxstep = c(4, 4, 12)){
cond <- make.dnf(cond, x)
if (cond %in% c("0", "1")) return(cond)
y <- as.vector(qcond_bool(cond,
ctInfo(configTable(x, rm.const.factors = FALSE, rm.dup.factors = FALSE, verbose = FALSE))\$scores))
if (isConstant(y)) return(as.character(y[[1]]))
x\$..RESP.. <- y
suppressMessages({
.cna <- cna(x, ordering = list("..RESP.."), strict = TRUE,
maxstep = maxstep, rm.const.factors = FALSE, rm.dup.factors = FALSE)
})
.asf <- asf(.cna)
if (attr(x, "type") == "mv"){
.asf <- subset(.asf, .asf\$outcome == "..RESP..=1")
}
lhs(.asf\$condition)
}
# Minimize multiple conditions
minimalize <- function(cond, x = NULL, maxstep = c(4, 4, 12)){
cond <- noblanks(cond)
if (is.null(x)){
x <- full.ct(cond)
} else {
x <- full.ct(x, cond = cond)
}
cti <- ctInfo(x)
out <- vector("list", length(cond))
names(out) <- cond
if (length(cond) == 0) return(out)
# check for disjunctive normal form
dnf <- checkValues(cond, c("+", "*"), colnames(cti\$scores))
cond1 <- cond
cond1[!dnf] <- vapply(cond[!dnf], make.dnf, x,
FUN.VALUE = character(1), USE.NAMES = FALSE)
out[] <- lapply(cond1, .minim1, x = x, maxstep = maxstep)
if (any(noOutput <- lengths(out) == 0))
warning("No minimal solution found for condition(s):\n",
paste0("  ", cond[noOutput], "\n"),
"You may try to increase maxstep.", call. = FALSE)
out
}
```

## Try the cna package in your browser

Any scripts or data that you put into this service are public.

cna documentation built on June 28, 2024, 5:08 p.m.