Aligned deep neural network for integrative analysis with high-dimensional input

Deep neural network (DNN) techniques have demonstrated significant advantages over regression and some other techniques. In recent studies, DNN-based analysis has been conducted on data with high-dimensional input such as omics measurements. In such analysis, regularization, in particular penalizati...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Journal of biomedical informatics Ročník 144; s. 104434
Hlavní autoři: Zhang, Shunqin, Zhang, Sanguo, Yi, Huangdi, Ma, Shuangge
Médium: Journal Article
Jazyk:angličtina
Vydáno: United States 01.08.2023
Témata:
Algorithms
Humans
Neoplasms
Neural Networks, Computer
Integrative analysis
Alignment
Penalization
DNN
High-dimensional
ISSN:1532-0480, 1532-0480
On-line přístup:Zjistit podrobnosti o přístupu
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Deep neural network (DNN) techniques have demonstrated significant advantages over regression and some other techniques. In recent studies, DNN-based analysis has been conducted on data with high-dimensional input such as omics measurements. In such analysis, regularization, in particular penalization, has been applied to regularize estimation and distinguish relevant input variables from irrelevant ones. A unique challenge arises from the "lack of information" attributable to high dimensionality of input and limited size of training data. For many data/studies, there exist other data/studies that may be relevant and can potentially provide additional information to boost performance. In this study, we conduct integrative analysis of multiple independent datasets/studies, with the goal of borrowing information across each other and improving overall performance. Significantly different from regression-based integrative analysis (where alignment can be easily achieved based on covariates), alignment across multiple DNNs can be nontrivial. We develop ANNI, an Aligned DNN technique for Integrative analysis with high-dimensional input. Penalization is applied for regularized estimation, selection of important input variables, and, equally importantly, information borrowing across multiple DNNs. An effective computational algorithm is developed. Extensive simulations demonstrate competitive performance of the proposed technique. The analysis of cancer omics data further establishes its practical utility.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:1532-0480
1532-0480
DOI:10.1016/j.jbi.2023.104434