fit_nonparaT: Main function for nonparametric tensor estimation and...

In TensorComplete: Tensor Noise Reduction and Completion Methods

Main function for nonparametric tensor estimation and completion based on low sign rank model.

Description

Estimate a signal tensor from a noisy and incomplete data tensor using nonparametric tensor method via sign series.

Usage

fit_nonparaT(Y,truer,H,Lmin,Lmax,option = 2)

Arguments

Y	A given (possibly noisy and incomplete) data tensor. The function allows both continuous- and binary-valued tensors. Missing value should be encoded as NA.
truer	Sign rank of the signal tensor.
H	Resolution parameter.
Lmin	Minimum value of the signal tensor (or minimum value of the tensor Y).
Lmax	Maximum value of the signal tensor (or maximum value of the tensor Y).
option	A large margin loss to be used. Use logistic loss if option = 1, hinge loss if option = 2. Hinge loss is default.

Value

The returned object is a list of components.

fitted - A series of optimizers that minimize the weighted classification loss at each level.

est - An estimated signal tensor based on nonparametic tensor method via sign series.

References

C. Lee and M. Wang. Beyond the Signs: Nonparametric Tensor Completion via Sign Series. Neural Information Processing Systems 34 (NeurIPS), 2021.

Examples

library(tensorregress)
indices = c(2,2,2)
noise = rand_tensor(indices)@data
Theta = array(runif(prod(indices),min=-1,max = 1),indices)

# The signal plus noise model
Y = Theta + noise

# Estimate Theta from nonparametic completion method via sign series
hatTheta = fit_nonparaT(Y,truer = 1,H = 1,Lmin = -1,Lmax = 1, option =2)
print(hatTheta$est)

TensorComplete documentation built on April 14, 2023, 9:10 a.m.