Skip to contents

Non-parametric regression

Source: R/further-regression.R
tda_npreg.Rd

tda_npreg is a non-parametric regression evaluated at points you choose: TDA's npreg, a kernel-smoothed mean, quantile or frequency.

Usage

tda_npreg(
  formula,
  data,
  x,
  method = c("mean", "quantile", "frequency", "lowess", "midmeans"),
  kernel = c("uniform", "triangle", "quartic", "epanechnikov"),
  bandwidth = NULL,
  select = NULL,
  options = list(),
  dir = tempfile("tda"),
  ...
)

Arguments

formula

a two-sided formula.

data

a data frame.

x

for tda_npreg, the points at which to evaluate the fit.

method

see Details; npreg's opt=. Either a name (partially matched) or TDA's number directly (opt=1 is "mean", and so on) – useful when translating a .cf file, which only ever writes the number.

kernel

the smoothing kernel, only meaningful for method = "mean" – npreg's k=. A name or TDA's number, the same as method.

bandwidth

the kernel bandwidth (npreg's d=, default 1).

select

a case-selection expression, TDA's sel=.

options

a named list of further TDA options, passed through.

dir

working directory.

...

passed to tda_run.

Value

An object carrying the fitted values as table.

Details

method picks which: "mean" (the default), a kernel-smoothed mean at each point in x; "quantile", smoothed quantiles; "frequency", smoothed frequencies. Only these three exist. npreg names the columns for method "mean" and "quantile", which tda_npreg reproduces; "frequency" writes columns that depend on the data (the distinct response categories) and are returned as TDA numbered them, V1, V2, ...

See also

Other regression: TDA_FAMILIES, TDA_QRMODELS, tda_freg(), tda_gdf(), tda_l1reg(), tda_lsreg(), tda_mlrc_design(), tda_mreg(), tda_nlreg(), tda_zreg()

Examples

set.seed(19)
d <- data.frame(x = round(rnorm(60), 3))
d$y <- round(1 + sin(d$x) + rnorm(60, sd = 0.2), 3)
np <- tda_npreg(y ~ x, d, x = seq(-2, 2, 0.5))
np$table
#>   RECN    X NX          XM        YM       YSD
#> 1    1 -2.0  3 -1.90300000 0.3163333 0.2787406
#> 2    2 -1.5  7 -1.27557143 0.1440000 0.1282433
#> 3    3 -1.0 15 -0.84100000 0.3114000 0.1956454
#> 4    4 -0.5 22 -0.45181818 0.6211364 0.3012182
#> 5    5  0.0 19 -0.07663158 0.9491053 0.3062905
#> 6    6  0.5 20  0.53270000 1.4328500 0.2978702
#> 7    7  1.0 17  0.84864706 1.6395882 0.2886109
#> 8    8  1.5  8  1.53475000 1.9345000 0.2688335
#> 9    9  2.0  5  1.83320000 1.9724000 0.2669191

# a triangle kernel, wider bandwidth
np2 <- tda_npreg(y ~ x, d, x = seq(-2, 2, 0.5), kernel = "triangle",
                 bandwidth = 0.5)
np2$table
#>   RECN    X NX          XM        YM       YSD
#> 1    1 -2.0  1 -1.82700000 0.1890000 0.0000000
#> 2    2 -1.5  1 -1.51300000 0.1240000 0.0000000
#> 3    3 -1.0  7 -1.07957143 0.1782073 0.1528711
#> 4    4 -0.5 14 -0.49385714 0.5344353 0.1960580
#> 5    5  0.0 11 -0.02218182 1.0768303 0.2161748
#> 6    6  0.5  9  0.51211111 1.4221763 0.2066423
#> 7    7  1.0  7  0.90300000 1.6029630 0.2058259
#> 8    8  1.5  5  1.47000000 1.9501246 0.1462970
#> 9    9  2.0  3  1.98566667 2.0862431 0.3072334