# Customizable sets Creation and manipulation of customizable sets. ```r cset(gset, orderfun = sets_options("orderfun"), matchfun = sets_options("matchfun")) cset_support(x) cset_core(x, na.rm = FALSE) cset_peak(x, na.rm = FALSE) cset_height(x, na.rm = FALSE) cset_memberships(x, filter = NULL) cset_universe(x) cset_bound(x) cset_transform_memberships(x, FUN, ...) cset_concentrate(x) cset_dilate(x) cset_normalize(x, height = 1) cset_defuzzify(x, method = c("meanofmax", "smallestofmax", "largestofmax", "centroid")) matchfun(FUN) cset_orderfun(x) cset_matchfun(x) cset_orderfun(x) <- value cset_matchfun(x) <- value as.cset(x) is.cset(x) cset_is_empty(x, na.rm = FALSE) cset_is_subset(x, y, na.rm = FALSE) cset_is_proper_subset(x, y, na.rm = FALSE) cset_is_equal(x, y, na.rm = FALSE) cset_contains_element(x, e) cset_is_set(x, na.rm = FALSE) cset_is_multiset(x, na.rm = FALSE) cset_is_fuzzy_set(x, na.rm = FALSE) cset_is_set_or_multiset(x, na.rm = FALSE) cset_is_set_or_fuzzy_set(x, na.rm = FALSE) cset_is_fuzzy_multiset(x) cset_is_crisp(x, na.rm = FALSE) cset_has_missings(x) cset_cardinality(x, type = c("absolute", "relative"), na.rm = FALSE) cset_union(...) cset_mean(x, y, type = c("arithmetic", "geometric", "harmonic")) cset_product(...) cset_difference(...) cset_intersection(...) cset_symdiff(...) cset_complement(x, y) cset_power(x) cset_cartesian(...) cset_combn(x, m) ## S3 method for class 'cset' cut(x, level = 1, type = c("alpha", "nu"), strict = FALSE, ...) ## S3 method for class 'cset' mean(x, ..., na.rm = FALSE) ## S3 method for class 'cset' ## median(x, na.rm = FALSE, ...) [R >= 3.4.0] ## median(x, na.rm) [R < 3.4.0] ## S3 method for class 'cset' length(x) ## S3 method for class 'cset' lengths(x, use.names = TRUE) ``` ## Arguments - `x`: For `as.cset()` and `is.cset()`: an object. A (c)set object otherwise. - `y`: A (c)set object. - `gset`: A generalized set (or some other object coercible to it). - `matchfun`: A function for matching similar elements, comparable to `match`, taking two arguments: `x` (vector of elements to be matched) and `table` (vector of elements to be matched against). The return value is an integer vector of the matching positions (or `NA` if there is no match). Note that the default behavior is to test for **identity**. - `FUN`: A predicate testing for equality of two objects. - `orderfun`: A function taking a list and returning an integer vector, specifying the order in which an iterator processes the set elements. Alternatively, the index vector can be specified directly. - `value`: A new match function (order function). - `type`: For `gset_cardinality()`: cardinality type (either `"absolute"` or `"relative"`). For `gset_mean()`: mean type (`"arithmetic"`, `"geometric"`, or `"harmonic"`). For `"cut"`: either `"alpha"` or `"nu"`. - `strict`: Logical indicating whether the cut level must be exceeded strictly (greater than ) or not (greater than or equal ). - `height`: Double from the unit interval for scaling memberships. - `e`: An object of class `element`. - `filter`: Optional vector of elements to be filtered. - `m`: Number of elements to choose. - `method`: Currently, only `"Jaccard"` is implemented. - `level`: The minimum membership level. - `use.names`: logical; should the names of `x` be used in the result? - `na.rm`: logical indicating whether `NA` values should be removed. - ``...``: For `cset_foo()`: (c)set objects. For the mean and sort methods: additional parameters internally passed to `mean` and `order`, respectively. For `gset_transform_memberships`: further arguments passed to `FUN`. For `cut`: currently not used. ## Details Customizable sets extend generalized sets in two ways: First, users can control the way elements are matched, i.e., define equivalence classes of elements. Second, an order function (or permutation index) can be specified for each set for changing the order in which iterators such as `as.list` process the elements. The latter in particular influences the labeling and print methods for customizable sets. The match function needs to be vectorized in a similar way than `match`. `matchfun` can be used to create such a function from a simple predicate testing for equality (such as, e.g., `identical`). Make sure, however, to create the same function only once. Note that operations on customizable sets require the same match function for all sets involved. The order function can differ, but will then be stripped from the result. `sets_options` can be used to conveniently switch the default match and/or order function if a number of `cset` objects need to be created. ## See Also `set` for (ordinary ) sets, `gset` for generalized sets, `cset_outer`, and `tuple` for tuples (vectors ). ## References D. Meyer and K. Hornik (2009), Generalized and customizable sets in R, **Journal of Statistical Software** 31 (2), 1--27. tools:::Rd_expr_doi("10.18637/jss.v031.i02") . ## Examples ```r ## default behavior of sets: matching of elements is very strict ## Note that on most systems, 3.3 - 2.2 != 1.1 x <- set("1", 1L, 1, 3.3 - 2.2, 1.1) print(x) y <- set(1, 1.1, 2L, "2") print(y) 1L %e% y set_union(x, y) set_intersection(x, y) set_complement(x, y) ## Now use the more sloppy match()-function. ## Note that 1 == "1" == 1L ... X <- cset(x, matchfun = match) print(X) Y <- cset(y, matchfun = match) print(Y) 1L %e% Y cset_union(X, Y) cset_intersection(X, Y) cset_complement(X, Y) ## Same using all.equal(). ## This is a non-vectorized predicate, so use matchfun ## to generate a vectorized version: FUN <- matchfun(function(x, y) isTRUE(all.equal(x, y))) X <- cset(x, matchfun = FUN) print(X) Y <- cset(y, matchfun = FUN) print(Y) 1L %e% Y cset_union(X, Y) cset_intersection(X, Y) cset_complement(X, Y) ### change default functions via set_option sets_options("matchfun", match) cset(x) cset(y) cset(1:3) <= cset(c(1,2,3)) ### restore package defaults sets_options("matchfun", NULL) ### customized order function FUN <- function(x) order(as.character(x), decreasing = TRUE) Z <- cset(letters[1:5], orderfun = FUN) print(Z) as.character(Z) ## converter for ordered factors keeps order o <- ordered(c("a", "b", "a"), levels = c("b", "a")) as.set(o) as.cset(o) ## converter for other data types keep order if the elements are unique: as.cset(c("A", "quick", "brown", "fox")) as.cset(c("A", "quick", "brown", "fox", "quick")) ```

cset function