interpolate.classes: Interpolate data between different classes.
In coffeemuggler/EMMAgeo: End-Member Modelling of Grain-Size Data

Description Usage Arguments Value Author(s) See Also Examples

This function interpolates grain-size data for different classes, either to higher or to lower resolution.

interpolate.classes(
  X,
  boundaries.in,
  boundaries.out,
  method = "natural",
  fixed.start = TRUE
)

`X`	`Numeric` matrix, input data set with m samples (rows) and n variables (columns).
`boundaries.in`	`Numeric` vector, class boundaries of the input data.
`boundaries.out`	`Numeric` vector, class boundaries of the output data.
`method`	`Logical` scalar, interpolation method, one out of "linear" (linear interpolation), "fmm" (cubic spline), "natural" (natural spline), "periodic" (periodic spline). Default is `"natural"`.
`fixed.start`	`Logocal` scalar, specifying if the outer boundaries should be set to the same values as in the original matrix, default is `TRUE`. This may become necessary to avoid interpolation errors, see example.

Numeric matrix, interpolated class values.

Michael Dietze, Elisabeth Dietze

EMMA, approx, spline

## load example data
data(example_X)
classes.in <- seq(from = 1, to = 10, length.out = ncol(X))
  
## Example 1 - decrease the class numbers
## define number of output classes
classes.out <- seq(1, 10, length.out = 20)

## interpolate the data set
Y <- interpolate.classes(X = X, 
                         boundaries.in = classes.in, 
                         boundaries.out = classes.out,
                         method = "linear")

## show original vs. interpolation for first 10 samples
plot(NA, xlim = c(1, 10), ylim = c(0, 40))
for(i in 1:10) {
  lines(classes.in, X[i,] * 20 + i)
  lines(classes.out, Y[i,] * 20 + i, col = 2)
}

## Example 2 - increase the class numbers
## define number of output classes
classes.out <- seq(1, 10, length.out = 200)

## interpolate the data set
Y <- interpolate.classes(X = X, 
                         boundaries.in = classes.in, 
                         boundaries.out = classes.out)

## show original vs. interpolation for first 10 samples
plot(NA, xlim = c(1, 10), ylim = c(0, 40))
for(i in 1:10) {
  lines(classes.in, X[i,] * 20 + i)
  lines(classes.out, Y[i,] * 20 + i, col = 2)
}