match_func: Estimate Inverse Mapping
In DMTL: Tools for Applying Distribution Mapping Based Transfer Learning

Description Usage Arguments Value Examples

This function estimates an inverse map g for a given set of knots (input) and values (output) corresponding to a certain map f i.e., given x, y | f: x --> y, match_func() estimates g: y --> x using linear interpolation.

1	match_func(knots, vals, new_vals, lims, get_func = FALSE)

`knots`	Vector containing knots for the distribution estimate.
`vals`	Vector containing distribution values corresponding to the knots.
`new_vals`	Vector containing distribution values for which the knots are unknown. If missing, `match_func()` simply returns the map function.
`lims`	Vector providing the range of the knot values for mapping. If missing, these values are estimated from the given knots.
`get_func`	Flag for returning the map function if `new_vals` is provided. If `TRUE`, `match_func()` returns a named list with two components- `mapped` and `func` (mapped knots for `new_vals` and the mapping function, respectively). Defaults to `FALSE`.

If new_vals is missing, a function performing interpolation (linear or constant) of the given data points.

If get_func = FALSE, a vector containing the matched knots that will produce new_vals for the map f.

If get_func = TRUE, a named list with two components- mapped and func (mapped knots for new_vals and the mapping function, respectively).

set.seed(654321)
x <- rnorm(100, 1, 0.5)
F <- ecdf(x)
fval <- F(x)
map <- match_func(knots = x, vals = fval)

x2 <- rnorm(20, 0.8, 0.5)
F2 <- ecdf(x2)
fval2 <- F2(x2)
matched <- match_func(knots = x, vals = fval, new_vals = fval2)

## Plot histograms...
opar <- par(mfrow = c(1, 3))
hist(x);    hist(x2);    hist(matched)
par(opar)              # Reset par