sigmasq_clust: Maximum likelihood estimate of residual variance
In larslau/pavpop: Phase and amplitude varying population pattern models

Computes the linearized likelihood given the residual around a given warp and the corresponding Jacobians.

1 2	sigmasq_clust(param, n_par, r, Zis, amp_cov, warp_cov, t, tw, observation_weights)

`param`	variance parameters.
`n_par`	vector consisting of number of variance parameters for each covariance function.
`r`	residual.
`Zis`	list of Jacobians in the warps of the mean function around the given warp.
`amp_cov`	function for generating amplitude covariance matrix.
`warp_cov`	function for generating warp covariance function
`t`	array of time variables corresponding to r.
`tw`	anchor points for warp variables.
`observation_weights`	vector of weights for the individual functional samples to be applied to the likelihood. This is useful for clustering analysis.