Robust Regularized Estimator (RegMCD) for location and inverse scatter
Computes the Robust Regularized Estimator for location and inverse scatter.
rrest(data, lambda=0.5, hp=0.75, thresh=0.0001, maxit=10, penalty="L2")
- Matrix or data.frame of observations
- Penalty parameter which controls the sparseness of the resulting inverse scatter matrix. Default is 0.5
- Robustness parameter which specifies the amount of observations to be included in the computations. Default is 0.75
- Threshold value controlling the convergence of the iterative algorithm. Default is 0.0001. In most cases this argument does not have to be supplied.
- Maximum number of iterations of the algorithm. Default is 10.
- Type of penalty to be applied. Possible values are "L1" and "L2".
The Robust Regularized Estimator computes a sparse inverse covariance matrix of the given observations by maximization of a penalized likelihood function. The sparseness is controlled by a penalty parameter lambda. Possible outliers are dealt with by a robustness parameter alpha which specifies the amount of observations for which the likelihood function is maximized.
- The resulting location estimate.
- The resulting inverse covariance estimate.
- An index vector specifying the data subset used (see robustness parameter alpha).
- The maximized objective value.
- The maximized (log-)likelihood value.
- The number of iterations
x <- cbind(rnorm(100), rnorm(100), rnorm(100)) # use first group only rr <- rrest(x, lambda=0.2, hp=0.75) solve(rr$covi) ## estimated cov matrix