estimation: Estimation of tensor entries from the cumulative model

In tensorordinal: Tensor Noise Reduction and Completion Based on Ordinal Observations

Description

Estimate the ordinal-valued tensor entries given latent parameters and a type of estimations

Usage

estimation(theta,omega,type = c("mode","mean","median"))

Arguments

theta	a continuous-valued tensor (latent parameters)
omega	the cut-off points
type	type of estimations: "mode" specifies argmax based label estimation "mean" specifies mean based label estimation "median" specifies median based label estimation

Value

an estimated ordinal tensor given latent parameters and a type of estimations

References

Lee, C., & Wang, M. (2020). Tensor denoising and completion based on ordinal observations. arXiv preprint arXiv:2002.06524.

Examples

indices <- c(10,20,30)
arr <- array(runif(prod(indices),-2,2),dim = indices)
b <- c(-1.5,0,1.5)
r_predict <- estimation(arr,b,type = "mode");r_predict

tensorordinal documentation built on Jan. 13, 2021, 9:17 p.m.