Maximum Smoothed Likelihood Density Estimation for Inverse Problems

Extended HALS algorithm for nonnegative Tucker decomposition and its applications for multiway analysis and classification

“…For example, by applying the ISRA (Image Space Reconstruction Algorithm) update rule [39][40][41][42], also often referred to as Lee-Seung algorithms [43], we obtain the following multiplicative update rule…”

PARAFAC algorithms for large-scale problems

“…Consider two data tensors Y (1) ∈ R I×J×K 1 , Y (2) ∈ R I×J×K 2 with their approximated PARAFAC models: (15) where…”

“…By replacing Y (1) and Y (2) in (20) by their approximations in (15), and setting (20) to zero, we obtain an up-…”

“…From the gradients (41), we apply a similar method to derive the multiplicative LS algorithm for NMF [12][13][14][15][16] to obtain the update rule…”

Nonparametric Estimation of a Mixing Density via the Kernel Method

“…However such methods are applicable essentially only in the errors in variables model, that is, when the differences Xi -Yi are independent of each other and of the Yi'S and have a common distribution. Other methods are motivated from computational considerations, such as a direct adaptation of the EM algorithm (Vardi and Lee 1993), stopping it before converging (Laird and Louis 1991) or smoothing each step (Eggermont andLaRiccia 1995, Silverman et al 1985).…”