compute_mvd: Computing multiview distance
In ong8181/macam: A collection of miscellaneous R functions for ecology

compute_mvd

R Documentation

Computing multiview distance

Description

compute_mvd Compute multiview distance

Usage

compute_mvd(
  block_mvd,
  effect_var,
  E,
  tp = 1,
  lib = c(1, nrow(block_mvd)),
  pred = lib,
  make_block_method = "naive",
  make_block_max_lag = E,
  n_ssr = 10000,
  k = floor(sqrt(n_ssr)),
  random_seed = 1234,
  distance_only = TRUE,
  silent = FALSE
)

Arguments

`block_mvd`	Data.frame contains time series data. The first column should be the target column.
`effect_var`	Character or Numeric. Column name or index of the effect variable.
`E`	Numeric. Optimal embedding dimension of `effect_var`
`tp`	Numeric. Forecasting time ahead.
`lib`	Numeric vector. Library indices.
`pred`	Numeric vector. Prediction indices.
`make_block_method`	Character. If `naive`, the multivariate data.frame (`block_mvd`) is directly used to calculate the multiview distance. If `rEDM`, `rEDM::make_block()` is used to add the time-delayed ordinate for each variable, which make the method equivalent to Chang et al. (2021).
`make_block_max_lag`	Numeric. `max_lag` in `rEDM::make_block()`. This argument will be used only if `make_block_method = "rEDM"`.
`n_ssr`	Numeric. The total number of embeddings examined.
`k`	Numeric. The number of embeddings used to calculate ensemble distance.
`random_seed`	Numeric. Random seed.
`distance_only`	Logical. if `TRUE`, only distance matrix is returned.
`silent`	Logical. if `TRUE`, progress message will not be shown.