bttdSsvdImpute_ALS: bttdSsvdImpute_ALS

In namgillee/BTTSoftImpute: R Software for Block Tensor Train Decomposition for Missing Data Estimation

Description

Sparse SVD Imputation based on block TT format. It uses ALS to optimize each TT cores(V_i) and left-singular-vectors(U), i.e., U -> V1 -> V2 -> ...

Usage

bttdSsvdImpute_ALS(Y, lambda, ttrank_max, tol_dx = -1, tol_df = -1,
  tol_f = 0, maxit, verbose, X0 = NULL, final.svd = FALSE)

Details

Note: In the version 'als', the matrix Afill = Ares + udv' are updated at every iteration of updating u and each core tensor V_n, i.e., U -> (update) -> V1 -> (update) -> V2 -> ... -> V1 -> (update).

For inputs, see bttdSoftImpute.R. It is assumed that X0 is not NULL but given.

Last modified: 2018.1.15. by Namgil Lee (Kangwon National University)

Value

X: LR format, with attribute 'obj'

namgillee/BTTSoftImpute documentation built on July 2, 2019, 8:35 p.m.