myALS_SVD {mwTensor}

Alternating Least Square Singular Value Decomposition (ALS-SVD) as an example of user-defined matrix decomposition.

Description

The input data is assumed to be a matrix. When algorithms of MWCAParams and CoupledMWCAParams are specified as "myALS_SVD", This function is called in MWCA and CoupledMWCA.

Usage

myALS_SVD(Xn, k, L2=1e-10, iter=30)

Arguments

Xn

The input matrix which has N-rows and M-columns.

k

The rank parameter (k <= min(N,M))

L2

The regularization parameter (Default: 1e-10)

iter

The number of iteration (Default: 30)

Value

The output matrix which has N-rows and k-columns.

Author(s)

Koki Tsuyuzaki

References

Madeleine Udell et al., (2016). Generalized Low Rank Models, Foundations and Trends in Machine Learning, 9(1).

Examples

  if(interactive()){
    # Test data
    matdata <- matrix(runif(10*20), nrow=10, ncol=20)
    # Perform ALS-SVD
    myALS_SVD(matdata, k=3, L2=0.1, iter=10)
  }