ptf: Fit a Probit Tensor Factorization Model

In ptf: Probit Tensor Factorization

Description

Fit a Probit Tensor Factorization Model

Usage

ptf(X, k, n, r = 0, max_iter = 1000, tol = 1e-08, tol_M = 1e-05,
  iter_M_max = 2, print_option = TRUE)

Arguments

X	response data, which is a three-way array of size n by n by k
k	number of relations
n	number of entities
r	decomposition rank
max_iter	max number of iterations
tol	tolerance of absolute change in likelihood
tol_M	tolerance of absolute change in the M step
iter_M_max	max number of iterations for M step
print_option	whether print loss for each iteration or not

Value

fitted parameters

References

@references Ye Liu, 2021. Computational Methods for Complex Models with Latent Structure. PhD thesis with link at https://repository.lib.ncsu.edu/bitstream/handle/1840.20/37507/etd.pdf?sequence=1

Examples

n <- 20
k <- 10
r <- 3
p <- c(n, n, k)
X <- array(rnorm(prod(p)),dim=p)
X_binary <- ifelse(X < -1.5,1,0)
X_binary_with_missing <- X_binary
num_missing <- 200
missing_index <- data.frame(x1=sample(1:n,num_missing,replace=TRUE), 
                           x2=sample(1:n,num_missing,replace=TRUE),  
                           x3=sample(1:k,num_missing,replace=TRUE))
for(i in 1:num_missing){
 X_binary[missing_index[i,1],
          missing_index[i,2],
          missing_index[i,3]] <- NA
}
result <- ptf(X_binary_with_missing,k,n,r,print_option=FALSE) 

ptf documentation built on June 15, 2021, 5:06 p.m.