smoothing.matrix: Optimal Smoothing Matrix
In arorar/covmat: Covariance Matrix Estimation

Description Usage Arguments Details Author(s)

Optimal Smoothing Matrix

1 2	smoothing.matrix(R, startup_period = 10, training_period = 60, seed = 9999, trials = 50, method = "L-BFGS-B", lambda = 0.2)

`R`	data
`startup_period`	length of samples required to calculate initial values
`training_period`	length of samples required to calculate forecast errors for evalualating the objective
`seed`	random seed to replicate the starting values for optimization
`trials`	number of strarting values to try for any optimization. Large number of trials for high dimensions can be time consuming
`method`	optimization method to use to evaluate an estimate of smoothing matrix. Default is L-BFGS-B
`lambda`	known constant as described in the paper. Defaulted to 0.2

Calcuation of smoothing matrix is done by assuming that the smoothing matrix is symmetrix and has a spectral decomposition. The orthogonal matrix in the decomposition is calculated using the product of givens rotation matrices and requires d(d-1)/2 angles for a d dimensional matrix. The eigenvalues are restricted to lie in [0,1].

Rohit Arora

arorar/covmat documentation built on May 10, 2019, 1:48 p.m.