Computes the covariance function (surface) for functional data.
For 1D: Cov(s, t) = E[(X(s) - mu(s))(X(t) - mu(t))]
For 2D: Covariance across the flattened domain.
Arguments
- fdataobj
An object of class 'fdata'.
- ...
Additional arguments (currently ignored).
Value
A list with components:
- cov
The covariance matrix (m x m for 1D, (m1m2) x (m1m2) for 2D)
- argvals
The evaluation points (same as input)
- mean
The mean function
Examples
# 1D functional data
t <- seq(0, 1, length.out = 50)
X <- matrix(0, 20, 50)
for (i in 1:20) X[i, ] <- sin(2*pi*t) + rnorm(50, sd = 0.2)
fd <- fdata(X, argvals = t)
cov_result <- cov(fd)
image(cov_result$cov, main = "Covariance Surface")
