ecomplex: Compute the epsilon-complexity of a time series.
In nateaff/ecomplex: Compute the epsilon-complexity of a time series

Compute the epsilon-complexity of a time series.

1
2
3

ecomplex(x, ds = 6, max_degree = 5, method = c("cspline", "bspline",
  "lift", "all"), err_norm = c("mae", "mse", "max"), sample_type = c("step",
  "random"))

`x`	A vector of points.
`ds`	Number of times to downsample the input sequence.
`max_degree`	The maximum order spline used in the approximation step
`method`	The interpolation or approximation method. One of c("bspline", "cspline")
`err_norm`	The norm type used in computing the approximation error.
`sample_type`	The downsampling type. Either randomly sampled or downsampled in integer steps.

A list with :

`A`	The epsilon-complexity intercept coefficient
`B`	The epsilon-complexity slope coefficient
`fit`	The full linear model generated by fitting log(epsilons) ~ log(S) using `lm()`.
`epsilons`	The mean sum of absolute errors at each downsample level.
`S`	The fraction of samples maintained at each downsample level.
`method`	The method used or a list of methods if method "all" is used.