Skip to Main content Skip to Navigation
Conference papers

A general iterative imputation scheme with feedback control for tensor completion (IFCTC)

Abstract : Tensors and tensor decompositions are very useful mathematical tools for representing and analyzing multidimensional data. The problem of estimating missing data in a tensor of measurements, named tensor completion, plays an important role in numerous applications. In this paper, to solve this problem, we propose a general iterative imputation scheme including a first-order feedback mechanism, aiming to improve algorithm performance. Two particularizations of this scheme, in which we apply soft and hard thresholding operators based on the Tucker model, are discussed. Then, simulation results are presented to illustrate their performance.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01590736
Contributor : José Henrique de Morais Goulart <>
Submitted on : Wednesday, September 20, 2017 - 10:13:01 AM
Last modification on : Thursday, August 6, 2020 - 3:59:29 AM

Identifiers

  • HAL Id : hal-01590736, version 1

Collections

Citation

José Henrique de Morais Goulart, Gérard Favier. A general iterative imputation scheme with feedback control for tensor completion (IFCTC). XXVIème colloque GRETSI (GRETSI 2017), Sep 2017, Juan-Les-Pins, France. ⟨hal-01590736⟩

Share

Metrics

Record views

394