TDbasedUFE

TDbasedUFE performs tensor decomposition-based unsupervised feature extraction to extract informative features from high-dimensional multiomics and bioinformatics datasets.

Key Features:

• Tensor decomposition-based unsupervised FE: Implements tensor decomposition for unsupervised feature extraction from multi-dimensional datasets.
• Multi-dimensional data handling: Extracts meaningful patterns and features from tensors without requiring prior labeling or supervision.
• Comparative performance: Demonstrated superior performance relative to DESeq2 and DIABLO for identifying differentially expressed genes and in multiomics analyses.
• Package variants: Distributed as two packages, TDbasedUFE and TDbasedUFEadv, addressing general and advanced analysis needs respectively.

Scientific Applications:

• Biomarker Identification: Extracts features from complex datasets to support identification of candidate biomarkers for disease.
• Gene Discovery: Supports discovery of genes implicated in disease processes through unsupervised feature extraction.
• Drug Repositioning: Facilitates detection of feature signatures that can be leveraged for identifying new uses of existing drugs.

Methodology:

Applies tensor decomposition techniques to perform unsupervised feature extraction on multi-dimensional (tensor) data, extracting patterns and features without supervision.