LearnNonparam (version 1.2.3)

pmt: Syntactic Sugar for Object Construction

Description

Construct test objects in a unified way.

Usage

pmt(key, ...)
pmts(
  which = c("all", "onesample", "twosample", "ksample", "multcomp", "paired", "rcbd",
    "association", "table")
)
define_pmt(
  statistic,
  inherit = c("twosample", "ksample", "paired", "rcbd", "association", "table"),
  rejection = c("lr", "l", "r"),
  scoring = c("none", "rank", "vw", "expon"),
  n_permu = 10000,
  name = "User-Defined Permutation Test",
  alternative = NULL,
  depends = character(),
  plugins = character(),
  includes = character()
)

Value

a test object corresponding to the specified key.

a data frame containing keys and corresponding tests implemented in this package.

Arguments

key

a character string specifying the test. Check pmts() for valid keys.

...

extra parameters passed to the constructor.

which

a character string specifying the desired tests.

statistic

definition of the test statistic. See Details.

inherit

a character string specifying the desired permutation test.

rejection

a character string specifying where the rejection region is.

scoring, n_permu

passed to the constructor.

name

a character string specifying the name of the test.

alternative

a character string specifying the alternative of the test.

depends, plugins, includes

passed to Rcpp::cppFunction().

Details

The test statistic in define_pmt can be defined using either R or Rcpp, with the statistic parameter specified as:

• R: a function returning a closure that returns a double.

• Rcpp: a character string defining a captureless lambda (introduced in C++11) returning another lambda that may capture by value, accepts const arguments of the same type, and returns a double.

When using Rcpp, the parameters for different inherit are listed as follows. Note that the parameter names are illustrative and may be modified.

• "twosample": (Rcpp::NumericVector sample_1, Rcpp::NumericVector sample_2)

• "ksample": (Rcpp::NumericVector combined_sample, Rcpp::IntegerVector one_based_group_index)

• "paired": (Rcpp::NumericVector sample_1, Rcpp::NumericVector sample_2)

• "rcbd": (Rcpp::NumericMatrix block_as_column_data)

• "association": (Rcpp::NumericVector sample_1, Rcpp::NumericVector sample_2)

• "table": (Rcpp::IntegerMatrix contingency_table)

Defining the test statistic using R follows a similar approach. The purpose of this design is to pre-calculate certain constants that remain invariant during permutation.

Examples

pmt("twosample.wilcoxon")

pmts()

# \donttest{
r <- define_pmt(
    inherit = "twosample", rejection = "lr", n_permu = 1e5,
    statistic = function(x, y) {
        m <- length(x)
        n <- length(y)
        function(x, y) sum(x) / m - sum(y) / n
    }
)

cpp <- define_pmt(
    inherit = "twosample", rejection = "lr", n_permu = 1e5,
    statistic = "[](NumericVector x, NumericVector y) {
        R_len_t n_x = x.size();
        R_len_t n_y = y.size();
        return [n_x, n_y](const NumericVector x, const NumericVector y) {
            double sum_x = 0;
            double sum_y = 0;
            for(auto x_i : x) sum_x += x_i;
            for(auto y_i : y) sum_y += y_i;
            return sum_x / n_x - sum_y / n_y;
        };
    }"
)

x <- rnorm(100)
y <- rnorm(100, 1)
options(LearnNonparam.pmt_progress = FALSE)
system.time(r$test(x, y))
system.time(cpp$test(x, y))
# }