rsvddpd 1.0

• Initial version of the package. Contains four functions.
  • rSVDdpd performs the robust svd using density power divergence. Also provides different convergence rule based warnings.
  • AddOutlier adds different types of outliers in a matrix.
  • simSVD simulates different scenarios and evaluates performance of SVD under pure model or under contamination.
  • cv.alpha function for finding the optimal robustness parameter.

rsvddpd 1.0.1

• Added a new function rank.rSVDdpd which is used to determine the rank for matrix factorization. This is internally applied to the existing function rSVDdpd for automatic matrix rank determination.